Search

With the popularity of the internet, the expression of emotions and methods of communication are becoming increasingly abundant, and most of these emotions are transmitted in text form. Text sentiment classification research mainly includes three methods based on sentiment dictionaries, machine learning and deep learning. In recent years, many deep learning-based works have used TextCNN (text convolution neural network) to extract text semantic information for text sentiment analysis. However, TextCNN only considers the length of the sentence when extracting semantic information. It ignores the semantic features between word vectors and only considers the maximum feature value of the feature image in the pooling layer without considering other information. Therefore, in this paper, we propose a convolutional neural network based on multiple convolutions and pooling for text sentiment classification (variable convolution and pooling convolution neural network, VCPCNN). There are three contributions in this paper. First, a multiconvolution and pooling neural network is proposed for the TextCNN network structure. Second, four convolution operations are introduced in the word embedding dimension or direction, which are helpful for mining the local features on the semantic dimensions of word vectors. Finally, average pooling is introduced in the pooling layer, which is beneficial for saving the important feature information of the extracted features. The verification test was carried out on four emotional datasets, including English emotional polarity, Chinese emotional polarity, Chinese subjective and objective emotion and Chinese multicategory. Our apporach is effective in that its result was up to 1.97% higher than that of the TextCNN network.

In the image forensics, detection of Cut-Paste manipulation is complicated computing. In this paper, the texture analysis of the spliced image is used to detect image forensics. From the local entropy of the median filter residual (MFR) of the forged image, the feature set is extracted for the ground truth mask `Find Gray level regional Maxima (FGM),' and `Entropy-based Edge (EbE).' Also, from the local range, the feature set is extracted for ground truth mask {`Morphological-Open Image (MOI), and 'Morphological- Erosion Image (MOE)'}. The feature vector in this paper composed of the two MOIs, two MOEs, and one EbE. The defined novel feature vector trained on a cubic support vector machine (SVM) classifier for only the performance evaluation of the proposed scheme. The performance of the proposed image forensics detection (IFD) scheme was measured with the five transformed types of image: median filtered (window size: {3×3, 5×5}), JPEG compressed (quality factor: {90, 70}) and average filtered (window size: 3×3). For the detection of the spliced image forensics, the region of Cut-Paste is classified by an input image only to the proposed scheme without the need for the trained SVM classifier. Throughout the experiment, the accuracy of median filtering detection was 98% over. Also, The area under the curve by sensitivity (TP: true positive rate) and 1-specificity (FP: false positive rate) results of the proposed IFD scheme approached to `1' with the trained cubic SVM classifier. Experimental results show high efficiency and performance to the spliced image. Therefore, the grade evaluation of the proposed scheme is ``Excellent (A)”.

Today’s rapidly developing communication technologies and dynamic collaborative business models made the security of data and resources more crucial than ever especially in multi-domain environments like Cloud and Cyber-Physical Systems (CPS). It enforced the research community to develop enhanced access control techniques and models for resources across multi-domain distributed environments so that the security requirements of all participating organizations can be fulfilled through considering dynamicity of changing environments and versatility of access control policies. The popularity of Role-Based Access Control (RBAC) model is irrefutable because of low administrative overhead and large-scale implementation in business organizations. However, it does not incorporate the dynamically changing policies and lacks semantically meaningful business roles which could have a diverse impact upon access decisions in multi-domain business environments. This paper describes our proposed novel access control framework that uses semantic business roles and intelligent agents through implementation of our Intelligent RBAC (I-RBAC) model. It encompasses occupational entitlements as roles for multiple domains. We use the dataset of original occupational roles provided by Standard Occupational Classification (SOC), USA. The novelty of the paper lies in developing a core I-RBAC ontology using real-world semantic business roles and intelligent agent technologies together for achieving required level of access control in highly dynamic multi-domain environment. The intelligent agents use WordNet and bidirectional LSTM deep neural network for automated population of organizational ontology from unstructured text policies. This dynamically learned organizational ontology is further matched with our core I-RBAC ontology in order to extract unified semantic business roles. The proposed I-RBAC model is mathematically described and the overall I-RBAC framework and its implementation architecture is explained. At the end, the I-RBAC model is validated through the implementation results that show a linear runtime trend of the model in presence of a large number of permission assignments and multiple queries.

According to the recent studies, malicious software (malware) is increasing at an alarming rate, and some malware can hide in the system by using different obfuscation techniques. In order to protect computer systems and the Internet from the malware, the malware needs to be detected before it affects a large number of systems. Recently, there have been made several studies on malware detection approaches. However, the detection of malware still remains problematic. Signature-based and heuristic-based detection approaches are fast and efficient to detect known malware, but especially signature-based detection approach has failed to detect unknown malware. On the other hand, behavior-based, model checking-based, and cloud-based approaches perform well for unknown and complicated malware; and deep learning-based, mobile devices-based, and IoT-based approaches also emerge to detect some portion of known and unknown malware. However, no approach can detect all malware in the wild. This shows that to build an effective method to detect malware is a very challenging task, and there is a huge gap for new studies and methods. This paper presents a detailed review on malware detection approaches and recent detection methods which use these approaches. Paper goal is to help researchers to have a general idea of the malware detection approaches, pros and cons of each detection approach, and methods that are used in these approaches.

The work in this paper proposes the application of the pinball quantile loss function to guide a deep neural network for Non-Intrusive Load Monitoring. The proposed architecture leverages concepts such as Convolution Neural Networks and Recurrent Neural Networks. For evaluation purposes, this paper also presents a set of complementary performance metrics for energy estimation. Finally, this paper also reports on the results of a comprehensive benchmark between the proposed network and three alternative deep neural networks, when guided by the pinball and Mean Squared Error loss functions. The obtained results confirm the disaggregation superiority of the proposed system, while also showing that the performances obtained using the pinball loss function are consistently superior to the ones obtained using the Mean Squared Error loss.

Financial risk management avoids losses and maximizes profits, and hence is vital to most businesses. As the task relies heavily on information-driven decision making, machine learning is a promising source for new methods and technologies. In recent years, we have seen increasing adoption of machine learning methods for various risk management tasks. Machine-learning researchers, however, often struggle to navigate the vast and complex domain knowledge and the fast-evolving literature. This paper fills this gap, by providing a systematic survey of the rapidly growing literature of machine learning research for financial risk management. The contributions of the paper are four-folds: First, we present a taxonomy of financial-risk-management tasks and connect them with relevant machine learning methods. Secondly, we highlight significant publications in the past decade. Thirdly, we identify major challenges being faced by researchers in this area. And finally, we point out emerging trends and promising research directions.

This paper proposed a novel first-order single-valued neutrosophic hesitant fuzzy time series (SVNHFTS) forecasting model. Our aim is to improve the previously proposed neutrosophic time series (NTS) model by incorporating the degree of the hesitancy using single-valued neutrosophic hesitant fuzzy set (SVNHFS) model instead of single-valued neutrosophic set (SVNS). Our paper's novelty is that we incorporate an algorithm that automatically converts the crisp dataset into the neutrosophic set that eliminates the need for experts' input or opinions in determining the membership in each of the partitioned neutrosophic set. We also incorporate Markov Chain algorithm in the de-neutrosophication process to include the weightage of the repeating neutrosophic logical relationships (NLRs). Our paper's significant contribution is to add to the existing body of knowledge related to fuzzy time series (FTS) by developing a new FTS model based on SVNHFS, one of the improved version of fuzzy sets, since this area of research is still relatively underdeveloped. To determine our proposed model's capability, we apply our proposed SVNHFTS model to three real datasets while also comparing the result to the other FTS models based on improved versions of fuzzy sets. Our datasets include benchmark enrollment data of University of Alabama, IDX Composite (Indonesian composite stock index), and MERVAL index (Argentinian composite stock index). The result shows that our proposed SVNHFTS model outperforms most of the other FTS models in terms of AFE and RMSE, especially the previously proposed NTS model.

Internet of Things (IoT) devices have connected millions of houses around the globe via the internet. In the recent past, threats due to hardware Trojan (HT) in the integrated circuits (IC) have become a serious concern, which affects IoT edge devices (IoT-ED). In this paper, the possibility of the IoT-ED with embedded HT that can cause serious security, privacy, and availability problems to the IoT based Home Area Network (HAN) has been discussed. Conventional network attack detection techniques work at the network protocol layers, whereas IoT-ED with HT can lead to the peculiar manifestation of attack at the physical and/or firmware level. On the other hand, in the IC design, most of the HT-based attack detection techniques require design time intervention, which is expensive for many of the IoT-ED and cannot guarantee 100% immunity. The argument in this paper is that the health of modern IoT-ED requires a final line of defense against possible HT-based attacks that goes undetected during IC design and test. The approach is to utilize power profiling (PP) and network traffic (NT) data without intervening into the IC design to detect malicious activity in HAN. The proposed technique is to effectively identify multiple attacks concurrently and to differentiate between different types of attacks. The IoT-ED behaviors for five different types of random attacks have been studied, including covert channel, DoS, ARQ, power depletion, and impersonation attacks. Data fusion has been leveraged by combining the PP and NT data and is able to detect, without design time intervention, each of the five attacks individually with up to 99% accuracy. Moreover, the proposed technique can also detect all the attacks concurrently with 92% accuracy. To the best of authors' knowledge, this is the first work where multiple HT based attacks are concurrently detected in IoT-ED without requiring any design time intervention.

A smart factory is a highly digitized and connected production facility that relies on smart manufacturing. Additionally, artificial intelligence is the core technology of smart factories. The use of machine learning and deep learning algorithms has produced fruitful results in many fields like image processing, speech recognition, fault detection, object detection, or medical sciences. With the increment in the use of smart machinery, the faults in the machinery equipment are expected to increase. Machinery fault detection and diagnosis through various deep learning algorithms has increased day by day. Many types of research have been done and published using both open-source and closed-source datasets, implementing the deep learning algorithms. Out of many publicly available datasets, Case Western Reserve University (CWRU) bearing dataset has been widely used to detect and diagnose machinery bearing fault and is accepted as a standard reference for validating the models. This paper summarizes the recent works which use the CWRU bearing dataset in machinery fault detection and diagnosis employing deep learning algorithms. We have reviewed the published works and presented the working algorithm, result, and other necessary details in this paper. This paper, we believe, can be of good help for future researchers to start their work on machinery fault detection and diagnosis using the CWRU dataset.

In this paper, a model for integrating production scheduling and maintenance planning is proposed for flow shop production system. The suggested model in this paper is based on the optimal jobs sequence for jobs that will be processed in multiple machines connected in series. The objective of this study is to find the optimal sequence for jobs while reducing the total production and maintenance costs. The model works by generating an initial solution using longest processing time (LPT) dispatching rule. Then, tabu search algorithm is established to obtain the optimal sequence for jobs. Computational experiments are performed on problems with five serially machines which are assigned to process eight diverse jobs from the same product family. The result is compared with the genetic algorithm optimization technique under individual PM scheme for obtaining superior solutions that has been proved in the literature to be one of the best approach. The computational results show that the recommended approach is qualified over the simulation based genetic algorithm optimization technique.

Extreme learning machine (ELM) is characterized by good generalization performance, fast training speed and less human intervention. With the explosion of large amount of data generated on the Internet, the learning algorithm in the single-machine environment cannot meet the huge memory consumption of matrix computing, so the implement of distributed ELM algorithm has gradually become one of the research focuses. In view of the research significance and implementation value of distributed ELM, this paper first introduced the research background of ELM and improved ELM. Secondly, this paper elaborated the implementation method of distributed ELM from the two directions: ensemble and matrix operation. Finally, we summarized the development status of distributed ELM and discussed the future research direction.

This paper proposes a semi-supervised autoencoder with an auxiliary task (SAAT) to extract a health feature space for power transformer fault diagnosis using dissolved gas analysis (DGA). The health feature space generated by a semi-supervised autoencoder (SSAE) not only identifies normal and thermal/electrical fault types, but also presents the underlying characteristics of DGA. In the proposed approach, by adding an auxiliary task that detects normal and fault states in the loss function of SSAE, the health feature space additionally enables visualization of health degradation properties. The overall procedure of the new approach includes three key steps: 1) preprocessing DGA data, 2) extracting two health features via SAAT, and 3) visualizing the two health features in two-dimensional space. In this paper, we test the proposed approach using massive unlabeled/labeled Korea Electric Power Corporation (KEPCO) databases and IEC TC 10 databases. To demonstrate the effectiveness of the proposed approach, four comparative studies are conducted with these datasets; the studies examined: 1) the effectiveness of an auxiliary detection task, 2) the effectiveness of the visualization method, 3) conventional fault diagnosis methods, and 4) the state-of-the-art, semi-supervised deep learning algorithms. By examining several evaluation metrics, these comparative studies confirm that the proposed approach outperforms SSAE without the auxiliary task, existing methods, and state-of-the-art deep learning algorithms, in terms of defining health degradation performance. We expect that the proposed SAAT-based health feature space approach will be widely applicable to intuitively monitor the health state of power transformers in the real world.

Safety analysis is often performed independent of the system design life cycle, leading to inconsistency between the system design and the safety artifact. Additionally, the process of generating safety artifacts is manual, time-consuming, and error-prone. As a result, safety analysis often requires re- work, is expensive, and increases system development time. Several model-based systems engineering (MBSE) approaches have been developed to automatically generate certain safety artifacts. However, these approaches only cover part of the system design and safety life cycle. To truly leverage the benefits of MBSE, system design must be undertaken together with safety analysis for the entire life cycle, and multiple safety artifacts must be generated from the same model. Moreover, MBSE approaches that require a model transformation between the system design and the safety model suffer from the inability to automatically reflect changes made to a safety artifact in the system and the safety model. This paper presents a framework to integrate the entire system design and safety life cycle using an MBSE approach. Both the system design and the safety data are captured in a single SysML model, from which safety artifacts such as failure modes and effects analysis (FMEA) tables and fault trees are automatically generated. This framework ensures consistency between the system design and the safety analysis by requiring no model transformation, thus reducing the resources required for safety analysis. The proposed Integrated System Design and Safety (ISDS) framework comprises three phases that together cover the entire system design and safety life cycle. In this paper, the application of Phase 1 of the framework to a real-world case study is demonstrated.

A fusion image combining both anatomical and functional information obtained by registering medical images of two different modalities, Positron Emission Tomography (PET) and Computed Tomography (CT), is of great significance for medical image analysis and diagnosis. Medical image registration relies on similarity measure which is low between PET/CT image voxels and therefore PET/CT registration is a challenging task. To address this issue, this paper presents an unsupervised end-to-end method, DenseRegNet, for deformable 3D PET/CT image registration. The method consists of two stages: (1) predicting 3D displacement vector field (DVF); and (2) registering 3D image. In the 3D DVF prediction stage, a two-level similarity measure together with a deformation regularization is proposed as loss function to optimize network training.In the image registration stage, a resampler and a spatial transformer are utilized to obtain the registration results. In this paper, 663 pairs of Uptake Value (SUV) and Hounsfield Unit (Hu) patches of 106 patients, 227 pairs of SUV and Hu patches of 35 patients and 259 pairs of SUV and Hu patches of 35 patients are randomly selected as training, validation and test set, respectively. Normalized cross correlation (NCC), intersection over union (IoU) of liver bounding box and euclidean distance (ED) on landmark points are used to evaluate the registration results. Experiment results show that the proposed method, DenseRegNet, achieves the best results in terms of liver bounding box IoU and ED, and the second highest value of NCC. For a trained model, given a new pair of PET/CT images, the registration result can be obtained with only one forward calculation within 10 seconds. Through qualitative and quantitative analyses, we demonstrate that, compared with other deep learning registration models, the proposed DenseRegNet achieves improved results in the challenging deformable PET/CT registration task.

The Generative Adversarial Networks (GAN) has been successfully applied to the generation of text content such as poetry and speech, and it is a hot topic in the field of text generation. However, GAN has been facing the problem of training and convergence. For the generation model, this paper redefines on the loss function. The truth-guided method has been added to make the generated text closer to the real data. For the discriminant model, this paper designs a more suitable network structure. The self-attention mechanism has been added to the discrimination network to obtain richer semantic information. Finally, some experiments under different model structures and different parameters indicates the model with truth-guided and self-attention mechanism gets better results.

This paper investigates the leader-following consensus tracking problems via iterative learning control for singular fraction-order multi-agent systems in the presence of iteration-varying switching topologies and initial state errors. First, in order to eliminate the impulsive effect of singular systems and handle iteration-varying topologies, the closed-loop Dα-type iterative learning control protocol is proposed. To deal with initial state errors, the initial state learning laws are introduced in light of the initial output errors of each follower agent. The developed Dα-type learning protocols based on initial state learning laws can guarantee each follower track perfectly the leader agent in the fixed time interval. Next, the sufficient convergent conditions of consensus tracking errors are provided. Moreover, the Dα-type learning protocols are extended to nonlinear singular fraction-order multi-agent systems with iteration-varying topologies and initial state errors. Finally, two numerical examples are presented to verify the validity of the proposed Dα type learning scheme in this paper.

Smart contracts, as an added functionality to blockchain, have received increased attention recently. They are executable programs whose instance and state are stored in blockchain. Hence, smart contracts and blockchain enable a trustable, trackable, and irreversible protocol without the need for trusted third parties which generally constitute a single point of failure. If a user creates and distributes a smart contract, others will be able to interact with it while the underlying blockchain ensures a trustable execution. In this paper, we aim to introduce state-of-the-art technologies of the smart contract protocol. We firstly introduce the history of blockchain and smart contracts followed by their step-by-step operations. Then, we introduce the survey results which are classified into four categories based on their purposes: cryptography, access management, social application, and smart contract structure. By presenting the most recent knowledge, this paper will contribute to the advances and proliferation of smart contracts.

In recent years, machine learning-based intrusion detection systems (IDSs) have proven to be effective; especially, deep neural networks improve the detection rates of intrusion detection models. However, as models become more and more complex, people can hardly get the explanations behind their decisions. At the same time, most of the works about model interpretation focuses on other fields like computer vision, natural language processing, and biology. This leads to the fact that in practical use, cybersecurity experts can hardly optimize their decisions according to the judgments of the model. To solve these issues, a framework is proposed in this paper to give an explanation for IDSs. This framework uses SHapley Additive exPlanations (SHAP), and combines local and global explanations to improve the interpretation of IDSs. The local explanations give the reasons why the model makes certain decisions on the specific input. The global explanations give the important features extracted from IDSs, present the relationships between the feature values and different types of attacks. At the same time, the interpretations between two different classifiers, one-vs-all classifier and multiclass classifier, are compared. NSL-KDD dataset is used to test the feasibility of the framework. The framework proposed in this paper leads to improve the transparency of any IDS, and helps the cybersecurity staff have a better understanding of IDSs' judgments. Furthermore, the different interpretations between different kinds of classifiers can also help security experts better design the structures of the IDSs. More importantly, this work is unique in the intrusion detection field, presenting the first use of the SHAP method to give explanations for IDSs.

Person re-identification, identifying and tracking pedestrians in cross-domain monitoring systems, is an important technology in the computer vision field and of real significance for the construction of smart cities. With the development of deep learning techniques, especially convolutional neural networks, this technology has received more extensive attention and improvement in recent years and a large number of noteworthy achievements have emerged. This paper provides a comprehensive overview of person re-identification approaches to assist researchers in quickly understand this field with preference as well as to provide a more structured framework. By reviewing more than 300 re-identification related papers, the focus of these studies is summarized as information extraction, metric learning, post-processing, efficiency improvement, labeling cost reduction, and data type expansion. This classification is then organized based on different technologies, and on this basis, the pros and cons of each technology are analyzed. Moreover, this overview summarizes the difficulties and challenges of re-identification and discusses the possible research directions for reference.

The Internet of Things (IoT) exemplifies a large network of sensing and actuating devices that have penetrated into the physical world enabling new applications like smart homes, intelligent transportation, smart healthcare and smart cities. Through IoT, these applications have consolidated in the modern world to generate, share, aggregate and analyze large amount of security-critical and privacy sensitive data. As this consolidation gets stronger, the need for security in IoT increases. With first line of defense strategies like cryptography being unsuited due to the resource constrained nature, second line of defense mechanisms are crucial to ensure security in IoT networks. This paper presents a comprehensive study of existing second line of defense mechanisms for standardized protocols in IoT networks. The paper analyzes existing mechanisms in three aspects: Intrusion Detection Systems (IDS), Intrusion Prevention Systems (IPS) and Intrusion Response Systems (IRS). We begin by providing an overview of standardized protocol stack, its layers and defensive security systems in IoT. From there, we build our narrative by presenting an extended taxonomy of IDS, IPS and IRS classifying them on their techniques, deployment, attacks, datasets, evaluation metrics and data pre-processing methods. We then thoroughly review, compare and analyze the research proposals in this context, considering the unique characteristics involved in these systems. Based on the extensive analysis of the existing defensive security systems, the paper also identifies open research challenges and directions for effective design of such systems for IoT networks, which could guide future research in the area.

The incorporation of the cloud technology with the Internet of Things (IoT) is significant in order to obtain better performance for a seamless, continuous, and ubiquitous framework. IoT has many applications in the healthcare sector, one of these applications is voice pathology monitoring. Unfortunately, voice pathology has not gained much attention, where there is an urgent need in this area due to the shortage of research and diagnosis of lethal diseases. Most of the researchers are focusing on the voice pathology and their finding is only to differentiating either the voice is normal (healthy) or pathological voice, where there is a lack of the current studies for detecting a certain disease such as laryngeal cancer. In this paper, we present an extensive review of the state-of-the-art techniques and studies of IoT frameworks and machine learning algorithms used in the healthcare in general and in the voice pathology surveillance systems in particular. Furthermore, this paper also presents applications, challenges and key issues of both IoT and machine learning algorithms in the healthcare. Finally, this paper highlights some open issues of IoT in healthcare that warrant further research and investigation in order to present an easy, comfortable and effective diagnosis and treatment of disease for both patients and doctors.

With the increased usage of Web 2.0 and data-affluent tools such as social media platforms and web blog services, the challenge of extracting public sentiment and disseminating personal health information has become more common than ever in the last decade. This paper proposes a novel model for Dengue disease detection based on social media posts alone. The model does not access any personal information of people or any medical record. The model extracts the presence of patients infected with Dengue disease based on tweets only and decides whether it is a general discussion about the disease, and no one is actually infected, or people are actually infected with that disease. The identification of people infected with Dengue is determined by clinical tests, but the propose technique is used for automatic surveillance and identification of regions where the spread is happening at an alarming rate and guide healthcare professional to take necessary actions to control the spread. This paper uses different machine/deep learning approaches to utilize tweets data for automatic and efficient disease detection. Experimental results demonstrate that the proposed model is able to achieve 92% accuracy compared to the current state-of-the-art techniques in this domain.

Notice of Violation of IEEE Publication Principles "Clustering Schemes in MANETs: Performance Evaluation, Open Challenges, and Proposed Solutions," by T. Rahman, I. Ullah, A. U. Rehman and R. A. Naqvi,in IEEE Access, vol. 8, January 2020, pp. 25135-25158 After careful and considered review of the content and authorship of this paper by a duly constituted expert committee, this paper has been found to be in violation of IEEE’s Publication Principles. This paper contains a figure (Figure 18) that was copied from the paper cited below. The figure was copied without attribution (including appropriate references to the original author(s) and/or paper title) and without permission. "BICSF: Bio-Inspired Clustering Scheme for FANETs," by A. Khan, F. Aftab and Z. Zhang,in IEEE Access, vol. 7, March 2019, pp. 31446-31456 Mobile ad hoc networks (MANETs) have recently received a lot of attention from researchers due to their ease of deployment and versatility. Clustering is one of the most efficient ways to coordinate routing in MANETs to improve network performance using different attributes and metrics. In this paper, the authors have extensively studied clustering schemes and divided the schemes into multiple types based on the Cluster Head (CH) selection criteria, which provides a good understanding of how each type of clustering algorithm differs from each other. The authors analyzed the performance of existing schemes based on the quality of service (QoS) metrics. Based on findings, the authors clarified some important tradeoffs between QoS metrics and also established some important factors influencing the efficiency of clustering schemes. These studies also contain open research challenges and proposed solutions to improve the performance of clustering schemes.

This paper proposes a neural generative architecture, namely NLDT, to generate a natural language short text describing a table which has formal structure and valuable information. Specifically, the architecture maps fields and values of a table to continuous vectors and then generates a natural language description by leveraging the semantics of a table. The NLDT architecture adopts a two-level neural model to make the most of the structure of a table to fully express the relationship between contents. To deal with the problem of out-of-vocabulary, this paper develops a simple and fast word-conversion method that replaces rare words appearing in texts with common field information in tables and directly replicates contents from table to the output sequence according to the field information. Besides, this paper adds the concept of theme to adapt the NLDT architecture to open domain and improves beam search algorithm to strengthen the results in the inference stage. On the WEATHERGOV dataset, the NLDT architecture improves the state-of-the-art BLEU-4 score from 61.01 to 62.89 and the current state-of-the-art F1 score from 73.21 to 78. On the WIKIBIO and WIKITABLE datasets, the NLDT architecture achieves a BLEU-4 score of 45.77 and 38.71 respectively which also outperform the state-of-the-art approaches. Furthermore, this paper introduces a Chinese dataset WIKIBIOCN including 33,244 biographies with corresponding tables. On the WIKIBIOCN dataset, the NLDT architecture achieves a BLEU-4 score of 38.87 and fairly good manual evaluation.

With the development of computer network technology and the expansion of network system, sensitive data is facing the threat of hacker attack. Intrusion detection is an active network security defense measure, which is an attempt to invade, an ongoing intrusion or an intrusion that has occurred to identify the process. At present, the detection rate of intrusion detection method is low, the false alarm rate and false alarm rate is high, and the real-time performance is poor. It needs a large number of or complete data to achieve better detection performance. In this paper, the concept, characteristics, classification, research contents and difficulties of traditional intrusion detection for mass multimedia data transmission network are described. Then, the basic principle of neural network and particle swarm optimization (PSO) algorithm and the basic idea of particle swarm optimization algorithm with quantum (QPSO) behaviour are introduced. It is emphasized that QPSO has better convergence performance than PSO algorithm in global optimization problems. In this paper, the concept, characteristics and structure of neural network are described, and the algorithm and classification of wavelet neural network are introduced. Then taking wavelet neural network (WNN) as the object, using the QPSO algorithm as the training algorithm, the concrete operation process is given. The research work in this paper shows that the performance of neural network trained by QPSO algorithm and improved QPSO algorithm is better than that of other intelligent algorithms such as PSO algorithm and genetic algorithm, and the convergence speed is faster than that of PSO algorithm or GA algorithm. QPSO is a high performance neural network training algorithm, which can play a good role in neural network anomaly detection.

The Internet of Things (IoT) introduces a new challenge for Database Management Systems (DBMS). In IoT, large numbers of sensors are used in daily lives. These sensors generate a huge amount of heterogeneous data that needs to be handled by the appropriate DBMS. The IoT has a challenge for the DBMS in evaluating how to store and manipulate a huge amount of heterogeneous data. DBMS can be categorized into two main types: The Relational DBMSs and the Non-relational DBMSs. This paper aims to provide a thorough comparative evaluation of two popular open-source DBMSs: MySQL as a Relational DBMS and MongoDB as a Non-relational DBMS. This comparison is based on evaluating the performance of inserting and retrieving a huge amount of IoT data and evaluating the performance of the two types of databases to work on resources with different specifications in cloud computing. This paper also proposes two prediction models and differentiates between them to estimate the response time in terms of the size of the database and the specifications of the cloud instance. These models help to select the appropriate DBMS to manage and store a certain size of data on an instance with particular specifications based on the estimated response time. The results indicate that MongoDB outperforms MySQL in terms of latency and the database size through increasing the amount of tested data. Moreover, MongoDB can save resources better than MySQL that needs resources with high capabilities to work with less performance.

This paper considers a time window periodic maintenance strategy with different duration windows and job scheduling activities in a single machine environment. The aim is to minimize the number of tardy jobs through the integration of production scheduling and periodic maintenance intervals. A mixed-integer linear programming model (MILP) is proposed to optimize small-sized test instances. Furthermore, an ant colony optimization (ACO) algorithm is developed to solve larger sized test instances. Subsequently, to measure the efficiency of the solutions obtained by ACO, Moore's algorithm is also developed to benchmark with ACO. To test the efficiency and the effectiveness of the ACO algorithm, a set of data for small and large sized problems was generated in which several parameters were adopted and then ten replicates were solved for each combination. The small sized instances were solved by the MILP. Then, the results obtained showed that the proposed ACO was able to obtain the exact solutions within reasonable CPU times, thus, it outperformed the CPLEX solver with respect to CPU. The large sized instances were solved by the Moore's algorithm and compared to ACO. Then, the results obtained showed that the ACO outperforms Moore's algorithm for all the instances tested. It can be concluded that the developed ACOis very efficient and effective in solving the problem considered in this paper.

In recent years, experts and scholars in the field of information security have attached great importance to the security of image information. They have proposed many image encryption algorithms with higher security. In order to further improve the security level of image encryption algorithm, this paper proposes a new image encryption algorithm based on two-dimensional Lorenz and Logistic. The encryption test of several classic images proves that the algorithm has high security and strong robustness. This paper also analyzes the security of encryption algorithms, such as analysis of the histogram, entropy process of information, examination of correlation, differential attack, key sensitivity test, secret key space analysis, noise attacks, contrast analysis. By comparing the image encryption algorithm proposed in this paper with some existing image encryption algorithms, the encryption algorithm has the characteristics of large secret key space, sensitivity to the key, small correlation coefficient and high contrast. In addition, the encryption algorithm is used. It can also resist noise attacks.

Intrusion detection technology, as an active and effective dynamic network defense technology, has rapidly become a hot research topic in the field of network security since it was proposed. However, current intrusion detection still faces some problems and challenges that affect its detection performance. Especially with the rapid development of the current network, the volume and dimension of network data are increasing day by day, and the network is full of a large number of unlabeled data, which brings great pressure on the data processing methods of IDS. In view of the tremendous pressure of intrusion detection brought by the current complex and high-dimensional network environment, this paper provides a feasible solution. Firstly, this paper briefly outlines the necessity of feature learning, the shortcomings of traditional feature learning methods and the new breakthroughs brought by deep belief network in feature learning, and focuses on the principle and working mechanism of deep belief network and Principal Component Analysis (PCA). Then, it constructs the intrusion detection model based on PCA-BP and DBN respectively. And through the experimental evaluation of the two detection models, a comparative experiment between deep belief network and principal component analysis is constructed. The experimental results show that deep belief network has unique advantages and good performance in feature learning. Therefore, deep belief network can be applied in the field of intrusion detection to extract effective features from the current high-dimensional and redundant network data, thereby improving the detection performance of IDS and its adaptability to the current complex and high-dimensional network environment.

Evolution of semiconductor manufacturing technology leads to the rising trend of leakage current and the end of Dennard scaling. At the dark silicon era, aggressive power gating scheme with quantitative management on power-gated hardware resources is required. This paper proposes a novel approach -parallelism scaling - to control static energy on power-gated parallel hardware. This work presents performance-constrained optimization method to power off the greatest possible amount of hardware. As a first trial, this paper examines the idea on VLIW-style architecture exploiting instruction-level parallelism. This paper establishes a theoretical foundation to realize parallelism scaling. The mathematical programming theory includes (1) topological model to control the granularity of program partitioning, (2) optimal partitioning on well-structured control flow graph in polynomial time, and (3) decision support for parallelism through item packing model guided by energy density. Evaluation conducted on EEMBC Denbench benchmark suite shows at least 15% to 53% saving on static energy compared to non-power-gated architecture. Compared to the state-of-art resource management scheme, our approach saves about 20% to 30% energy to meet the same performance demand. The evaluation reveals noteworthy opportunity to save static energy for future dark-silicon architecture design.

Applications of Blockchain (BC) technology and Cyber-Physical Systems (CPS) are increasing exponentially. However, framing resilient and correct smart contracts (SCs) for these smart application is a quite challenging task because of the complexity associated with them. SC is modernizing the traditional industrial, technical, and business processes. It is self-executable, self-verifiable, and embedded into the BC that eliminates the need for trusted third-party systems, which ultimately saves administration as well as service costs. It also improves system efficiency and reduces the associated security risks. However, SCs are well encouraging the new technological reforms in Industry 4.0, but still, various security and privacy challenges need to be addressed. In this paper, a survey on SC security vulnerabilities in the software code that can be easily hacked by a malicious user or may compromise the entire BC network is presented. As per the literature, the challenges related to SC security and privacy are not explored much by the authors around the world. From the existing proposals, it has been observed that designing a complex SCs cannot mitigate its privacy and security issues. So, this paper investigates various Artificial Intelligence (AI) techniques and tools for SC privacy protection. Then, open issues and challenges for AI-based SC are analyzed. Finally, a case study of retail marketing is presented, which uses AI and SC to preserve its security and privacy.

The cloud computing environment provides easy-to-access service for private and confidential data. However, there are many threats to the leakage of private data. This paper focuses on investigating the vulnerabilities of cloud service providers (CSPs) from three risk aspects: management risks, law risks, and technology risks. Additionally, this paper presents a risk assessment model that is based on grey system theory (GST), defines indicators for assessment, and fully utilizes the analytic hierarchy process (AHP). Furthermore, we use the GST to predict the risk values by using the MATLAB platform. The GST determines the bottom evaluation sequence, while the AHP calculates the index weights. Based on the GST and the AHP, layer-based assessment values are determined for the bottom evaluation sequence and the index weights. The combination of AHP and GST aims to obtain systematic and structured well-defined procedures that are based on step-by-step processes. The AHP and GST methods are applied successfully to handle any risk assessment problem of the CSP. Furthermore, substantial challenges are encountered in determining the CSP's response time and identifying the most suitable solution out of a specified series of solutions. This issue has been handled using two additive features: the response time and the grey incidence. The final risk values are calculated and can be used for prediction by utilizing the enhanced grey model (EGM) (1,1), which reduces the prediction error by providing direct forecast to avoid the iterative prediction shortcoming of standard GM (1,1). Thus, EGM (1,1) helps maintain the reliability on a larger scale despite utilizing more prediction periods. Based on the experimental results, we evaluate the validity, accuracy, and response time of the proposed approach. The simulation experiments were conducted to validate the suitability of the proposed model. The simulation results demonstrate that our risk assessment model contributes to reducing deviation to support CSPs with the three adopted models.

Virtual chemistry experiment is an important teaching tool in middle schools. It can not only help students understand the experimental principles, but also help them memorize the experimental procedures. However, there are still some problems in virtual chemistry experiments. First, in existing research, user intentions are often misunderstood and cannot be accurately understood. Second, the existing research fail to identify the user's wrong actions, which reduce the accuracy of the experiment. Third, the user sense of operation and realism is not strong during the experiment, which reduces the user's experience. Finally, the lack of navigation guidance for experimental operations increases user learning time. In order to solve these problems. This paper proposes a scheme for establishing an intelligent navigational chemical laboratory based on multimodal fusion. First of all, we design a new smart beaker structure with perceptual ability, which can be used to complete most chemical experiments and give users a real sense of experience. Besides, we propose a multimodal fusion understanding algorithm, which reduces the misidentification of the experiment and better understands the real intention of the user. Finally, intelligent navigation and wrong behavior recognition functions are added to the experimental equipment, which improves the efficiency of human-computer interaction. The results show that Compared with the existing virtual laboratory or system, the chemical laboratory scheme proposed in this paper through multimodal fusion understanding algorithm greatly reduces the user's memory load and improves the success rate of the experiment. Moreover, through the combination of virtual and real, the virtual chemistry experiment not only improves the authenticity of the operation, but also stimulates the students' interest in learning, which is well received by users.

The development of sensor technology provides massive data for data-driven fault diagnosis. In recent years, more and more scholars are studying artificial intelligence technology to solve the bottleneck in fault diagnosis. Compared with other classification and prediction problems, fault diagnosis often faces the problem of data scarcity. To overcome the lack of fault data, the transfer learning based on different working condition is gradually introduced into fault diagnosis by scholars. This paper discusses the current mainstream AI-based fault diagnosis methods, and analyzes the advantage of transfer learning for fault diagnosis problem. Then, a transfer component analysis (TCA) based method is proposed to transfer data features between different working conditions. Through the TCA-based method, the fault diagnosis model under the working condition can be established with the help of historical working condition. It effectively alleviates the problem of data scarcity under the condition to be predicted. Different from other fault diagnosis studies, this paper considers the online maintenance process based on TCA. A fault diagnosis framework including online maintenance process is proposed. Finally, a case study of bearing diagnosis from Case Western Reserve University proves the feasibility and effectiveness of the proposed TCA-based method and our fault diagnosis framework.

Most current multi-optimization survey papers classify methods into broad objective categories and do not draw clear boundaries between the specific techniques employed by these methods. This may lead to the misclassification of unrelated methods/techniques into the same objective category. Moreover, most of these survey papers classify algorithms as independent of the specific techniques they employ. Toward this end, we introduce in this survey paper a methodology-based taxonomy that classifies multi-optimization methods into hierarchically nested, fine-grained, and specific classes. We provide a methodological taxonomy to classify methods into the following hierarchical fashion: objective categories→ objective functions→optimization methods→optimization sub-methods. We introduce a comprehensive survey on the methods that are contained under each optimization method, the optimization methods contained under each objective function, and objective functions contained under each objective category. We selected the objective functions that should be maximized for solving most real-word multi-objective optimization problems, which are pairs of the following: partitions separability, internal density, dynamic similarity, and structural similarity. For each optimization method, we surveyed the various algorithms in literature that pertain to the method. We experimentally compared and ranked the optimization methods that fall under each objective function, the objective functions that fall under each objective category, and the objective categories used for solving a specific optimization problem.

The growing interest and recent breakthroughs in artificial intelligence and machine learning (ML) have actively contributed to an increase in research and development of new methods to estimate the states of electrified vehicle batteries. Data-driven approaches, such as ML, are becoming more popular for estimating the state of charge (SOC) and state of health (SOH) due to greater availability of battery data and improved computing power capabilities. This paper provides a survey of battery state estimation methods based on ML approaches such as feedforward neural networks (FNNs), recurrent neural networks (RNNs), support vector machines (SVM), radial basis functions (RBF), and Hamming networks. Comparisons between methods are shown in terms of data quality, inputs and outputs, test conditions, battery types, and stated accuracy to give readers a bigger picture view of the ML landscape for SOC and SOH estimation. Additionally, to provide insight into how to best approach with the comparison of different neural network structures, an FNN and long short-term memory (LSTM) RNN are trained fifty times each for 3000 epochs. The error is somewhat different for each training repetition due to the random initial values of the trainable parameters, demonstrating that it is important to train networks multiple times to achieve the best result. Furthermore, it is recommended that when performing a comparison among estimation techniques such as those presented in this review paper, the compared networks should have a similar number of learnable parameters and be trained and tested with identical data. Otherwise, it is difficult to make a general conclusion regarding the quality of a given estimation technique.

In the last few years, most of the public universities in the Kingdom of Saudi Arabia (KSA) have adopted electronic learning (e-learning) as an evolutionary educational platform both by choice and by demand. Despite its many advantages, it has increased the ethical and moral challenges faced in the form of the violations of academic integrity that exist in traditional education. Universities all over the world are struggling to overcome these issues. Keeping this in mind, the purpose of this paper is to investigate and priorities the factors of violating academic integrity within Saudi Arabian universities. Based on the previously published literature and using Delphi - based discussions with e-Learning experts, this paper has identified twelve major factors which were related to e-Learning environment, Awareness on Academic Integrity, and e-Learning guiding principles. Analytic Hierarchy Process (AHP) was used to prioritize these factors. It was concluded that inappropriate guidelines provided to students was the most contributing factor while Shortage of feedbacks was least contributing factor among all the identified factors.

As an essential and challenging technology of fault prediction and health management(PHM), fault prediction technology has been a research focus in the field of fault diagnosis. However, the current model-based fault prediction technology and data-driven fault prediction technology have some limitations, and it is difficult to effectively apply them in practice. Therefore, this paper combines the advantages of two kinds of fault prediction technology, sets the fault distribution function as the membership function of the adaptive fuzzy neural network based on the full analysis of the fault mechanism. The use of the fault distribution function to highly generalize the law of fault occurrence, and the strong self-learning ability of the neural network can effectively tap the potential fault information of the fault data, thereby using the fault distribution function to fit the fault data, and forming a set of membership functions by presetting a variety of membership functions, so as to expand the applicability of the proposed model in fault prediction. The experimental results show that the fault prediction model proposed in this paper has the advantages of high prediction accuracy, fast convergence speed and good applicability.

With the development of the Internet of Things (IoT) technology, many end-users participate in the smart city through their own intelligent mobile devices (such as personal wearable devices, smartphones.) or sensors. The main challenge of the device sensing layer in the edge computing system of the IoT in the smart city is to select the trusted participants. Because not all the intelligent devices of the IoT are trustworthy, some intelligent devices of the IoT may maliciously damage the network or services and affect the service quality of the system. On this basis, an intelligent device selective recommendation mechanism based on the dynamic black-and-white list was proposed to solve the problem of selecting trusted participants to improve the service quality of the edge computing system of the IoT in the smart city. We introduced the evolutionary game theory to theoretically qualitatively study the validity and stability of the trust management mechanism proposed in this paper. The Lyapunov theory was used to prove the validity and stability of the trust management mechanism. The effectiveness of the trust management mechanism was verified by the actual scenario of the personal health monitoring management system and the air-quality monitoring and analysis system in the smart city environment. Experiments showed that the trust management mechanism proposed in this paper has a significant role in promoting the cooperation of multi intelligent devices in the IoT edge computing system. It more reliably resists the malicious attacks to service providers and is suitable for the large-scale IoT edge computing system in the smart city.

The need for cybersecurity increases to protect the exchange of information for improving the data privacy. This paper presents an investigation of the encryption efficiency of the chaotic-based image block ciphering in the spatial and Fractional Fourier Transform (FrFT) domains. The main aim of this investigation is to examine the efficiency of different chaotic maps, while considering the parameters of the FrFT as additional keys for encryption and achieving reliable cybersecurity for robust image communication. In this paper, Cat, Baker, and Logistic map confusion approaches are applied in the spatial and FrFT domains to study and analyze the cybersecurity and ciphering efficiency of chaos-based image cryptosystems. The confusion features of the chaotic maps in spatial and FrFT domains are investigated using information entropy, differential analysis, histograms, visual observation, attack analysis, impact of noise, and encryption quality tests. Simulation results prove that the chaotic-based image encryption in the FrFT domain increases the efficiency of the confusion process and achieves a high nonlinear relation between the plainimage and the cipherimage in a symmetric ciphering approach. Moreover, the results demonstrate that the Cat-FrFT scheme is more susceptible to channel noise attacks than the Baker-FrFT and the Logistic-FrFT schemes. Hence, they can be implemented efficiently in the scenarios of noisy channels due to their high robustness to channel noise.

To improve the prediction effect of time series, we make a systematic study of various time series prediction methods based on statistics and machine learning in this paper. In the experiment, we compare the prediction results of several prediction methods. In particular, much research has been done on the selection of experimental data because representative time series data can better test the effectiveness and practicability of the prediction method. Based on the idea of divide and conquer of complex problems and the strategy of continuous optimization of machine learning, we proposed the prediction methods of LSTM-TFE, LR-TFE, and BR-TFE combined the EEMD, LSTM, LR, and BR methods in this paper. These methods use EEMD to decompose complex time series into several relatively milder, more regular and stable subsequences. Then the prediction model of each subsequence based on machine learning is carried out by using the LSTM, LR, or BR methods. We use these prediction models to predict the value of each subsequence. Finally, the value of multiple subsequences is fused to form the prediction results of the original complex time series. To verify the proposed method comprehensively, we select three representative time series data to test this paper. From the experimental results, we found that the proposed method has a good effect.

Portfolio optimization is a hot research topic, which has attracted many researchers in recent decades. Better portfolio optimization model can help investors earn more stable profits. This paper uses three deep neural networks (DNNs), i.e., deep multilayer perceptron (DMLP), long short memory (LSTM) neural network and convolutional neural network (CNN) to build prediction-based portfolio optimization models which own the advantages of both deep learning technology and modern portfolio theory. These models first use DNNs to predict each stock's future return. Then, predictive errors of DNNs are applied to measure the risk of each stock. Next, the portfolio optimization models are built by integrating the predictive returns and semi-absolute deviation of predictive errors. These models are compared with three equal weighted portfolios, where their stocks are selected by DMLP, LSTM neural network and CNN respectively. Also, two prediction-based portfolio models built with support vector regression are used as benchmarks. This paper applies component stocks of China securities 100 index in Chinese stock market as experimental data. Experimental results present that the prediction-based portfolio model based on DMLP performs the best among these models under different desired portfolio returns, and high desired portfolio return can further improve the performance of this model. This paper presents the promising performance of DNNs in building prediction-based portfolio models.

Recently, the ever-increasing vehicle population has become a severe challenge to traffic safety, especially the problem of a single-vehicle overtaking a platoon on the Two-Lane Two-Way (TLTW) road. Platooning has the potential to improve traffic efficiency and safety. However, there exists a perilous situation of “Neither overtake nor give up” when the single-vehicle overtakes a platoon on the TLTW road. This paper presents a flexible framework to automatically filter a large quantity of Advanced Driver Assistance Strategies (ADAS) and select the most suitable driver assistance information for the single-vehicle overtakes a platoon on the TLTW road. A step-by-step Single Vehicle Overtakes Platoon (SVOP) algorithm is designed to generate the coarse ADAS, which had given plenty of consideration to the vehicle safety, traffic efficiency, and driving comfort. Then, this paper obtains the raw data about the single-vehicle overtakes a platoon on the TLTW by using CARLA, which can help us to get 20 drivers' upper and down boundaries of both velocity and acceleration. In addition, the extracted ranges of velocity and acceleration are used to quantitatively analyse the drivers' driving features and filter the ADAS information. Finally, a Bayesian nonparametric approach is developed to segment driver's driving raw data temporal sequences into small analytically interpretable components without using prior knowledge. So that the accurately overtaking characteristics can be obtained, and the ADAS can be further filtered. Experimental results demonstrate that the obtained coarse ADAS are only valid in theory but not acceptable by most of the drivers. Nonetheless, by leveraging the nonparametric Bayes algorithm, the driver's overtaking behavior can be divided into different primitives, from which some could obtain the driver's acceptance range for the velocity and acceleration. 92.3% ~ 94.78% invalid SVOP ADAS could be filtered out by leveraging the primitive-based SVOP approach. Thus, after filtering, the overtaking scheme is the most acceptable strategy for drivers.

Search is an essential technology for rescue and other mobile robot applications. Many robotic search and rescue systems rely on teleoperation. One of the key problems in search tasks is how to cover the search space efficiently. Search is also central to humans’ daily activities. This paper analyzes and models human search behavior using data from actual teleoperation experiments. The analysis of the experimental data uses a novel technique to decompose search data, based on structure learning and K-means clustering. The analysis explores three hypotheses: (1) humans are able to solve a complex search task by breaking it up into smaller tasks, (2) humans consider both coverage and motion cost, and (3) robots can outperform humans in search problems. The enhanced understanding of human search strategies can then be applied to the design of human-robot interfaces and search algorithms. The paper describes a technique for augmenting human search. Since the objective functions in search problems are submodular, greedy algorithms can generate near-optimal subgoals. These subgoals then can be used to guide humans in searching. Experiments showed that the humans’ search performance is improved with the subgoals’ assistance.

Nowadays, Heart disease is one of the crucial impacts of mortality in the country. In clinical data analysis, predicting cardiovascular disease is a primary challenge. Deep learning (DL) has been demonstrated to be effective in helping to determine and forecast a huge amount of data produced by the health industry. In this paper, the proposed Recursion enhanced random forest with an improved linear model (RFRF-ILM) to detect heart disease. This paper aims to find the key features of the prediction of cardiovascular diseases through the use of machine learning techniques. The prediction model is adding various combinations of features and various established methods of classification. it produces a better level of performance with precision through the heart disease prediction model. In this study, the factors leading to cardiovascular disease can be diagnosed. A comparison of important variables showed with the Internet of Medical Things (IoMT) platform, for data analysis. This indicates that coronary artery disease develops more often in older ages. Also important in this disease's outbreak is high blood pressure. For this purpose, measures must be taken to prevent this disease and Diabetes provides a further aspect that should be taken into consideration in the occurrence of coronary artery disease with 96.6 % accuracy,96.8% stability ratio and 96.7% F-measure ratio.

With the development of medical digitization technology, artificial intelligence and big data technology, the medical model is gradually changing from treatment-oriented to prevention-oriented. In recent years, with the rise of artificial neural networks, especially deep learning, great achievements have been made in realizing image classification, natural language processing, text processing and other fields. Combining artificial intelligence and big data technology for disease risk prediction is a research focus in the field of intelligent medicine. Blood lipids are the main risk factors of cardiovascular and cerebrovascular diseases. If early prediction of abnormal blood lipids in iron and steel workers can be carried out, early intervention can be carried out, which is beneficial to protect the health of iron and steel workers. This paper around the steel workers dyslipidemia prediction problem for further study, firstly analyzes the influence factors of the steel workers dyslipidemia, discusses the commonly used method for prediction of disease, and then studied deep learning related theory, this paper introduces the two deep learning algorithms of RNN (Recurrent Neural Network) and LSTM (Long Short-Term Memory). Use the basic principle of Python language and the TensorFlow deep learning framework, establishes a prediction model based on two deep learning networks, and makes an example analysis. Experimental results show the LSTM prediction effect is superior to traditional RNN network, It provides scientific basis for the prevention of iron and steel dyslipidemia.

The detection of abnormal electricity consumption behavior has been of great importance in recent years. However, existing research often focuses on algorithm improvement and ignores the process of obtaining features. The optimal feature set, which reflects customers’ electricity consumption behavior, has a significant influence on the final detection results. Moreover, it is not straightforward to obtain datasets with label information. In this paper, a method based on feature engineering for unsupervised detection of abnormal electricity consumption behavior is proposed. First, the original feature set is constructed by brainstorming in the feature engineering step. Then, the optimal feature set, which reflects the customers’ electricity consumption behavior, is obtained by features selected based on the variance and similarity between them. After that, in the abnormal detection step, a density-based clustering algorithm, in which the best clustering parameters are selected through iteration and evaluation, combined with unsupervised clustering evaluation indexes, is used to detect abnormal electricity consumption behaviors. Finally, using the load dataset of an industrial park, several typical feature strategies are applied for comparison with the feature engineering proposed in this paper. To perform the evaluation, the label information of abnormal behaviors is obtained by combining the original electricity consumption behavior detection results with abnormal data injections. The abnormal detection method proposed has given good results and outperformed typical feature strategies in an effective and generalizable way.

NLP (Natural Language Processing) is a technology that enables computers to understand human languages. Deep-level grammatical and semantic analysis usually uses words as the basic unit, and word segmentation is usually the primary task of NLP. In order to solve the practical problem of huge structural differences between different data modalities in a multi-modal environment and traditional machine learning methods cannot be directly applied, this paper introduces the feature extraction method of deep learning and applies the ideas of deep learning to multi-modal feature extraction. This paper proposes a multi-modal neural network. For each mode, there is a multilayer sub-neural network with an independent structure corresponding to it. It is used to convert the features in different modes to the same-modal features. In terms of word segmentation processing, in view of the problems that existing word segmentation methods can hardly guarantee long-term dependency of text semantics and long training prediction time, a hybrid network English word segmentation processing method is proposed. This method applies BI-GRU (Bidirectional Gated Recurrent Unit) to English word segmentation, and uses the CRF (Conditional Random Field) model to annotate sentences in sequence, effectively solving the long-distance dependency of text semantics, shortening network training and predicted time. Experiments show that the processing effect of this method on word segmentation is similar to that of BI-LSTM-CRF (Bidirectional- Long Short Term Memory-Conditional Random Field) model, but the average predicted processing speed is 1.94 times that of BI-LSTM-CRF, effectively improving the efficiency of word segmentation processing.

In the current electrical load profile analysis, considering the shortage of traditional methods on the typical load profile extraction of single consumers and the load profile feature extraction, this paper proposes an approach based on time series data mining. Firstly, this method reduces the dimension of the load profile of a single consumer based on the Piecewise Aggregate Approximation(PAA), and re-expresses the load profile of the consumer over a period based on the Symbolic Aggregate approXimation(SAX), representing the consumer’s load profile with a symbolic string to extract the typical load profile. Then, combined with the load characteristic indices and time series-based features, the typical load profiles of different consumers are clustered based on the K-means algorithm to analyze the power consumption behaviors. Finally, this paper performs a case analysis with a UCI test data set, and the results show that the proposed approach can excavate typical power consumption behaviors of consumers and improve the electrical load profile analysis efficiency and the clustering quality.

Recently, high utility pattern mining (HUPM) is one of the most important research issues in data mining. Because it can consider the non-binary frequency values of items in a transaction and the different profit values of each item. It has been widely used. First of all, this paper briefly describes the related concepts, formulas and examples of application for HUPM. Secondly, the key technologies for HUMP are introduced in detail, and they are divided into main methods including Apriori-based, tree-based, projection-based, list-based, data format-based, and index-based and so on. The paper further compares data sets, uses, advantages and disadvantages of algorithms, laid the foundation for the next research direction. Then, this article outlines the high utility derivative patterns, including high average utility pattern, high utility sequential pattern, and high utility compact pattern and so on. Because static data is difficult to meet the actual needs, this paper summarizes the efficient use of HUPMs' methods over data streams, mainly based on incremental methods, based on the sliding window model methods, based on the time decay model methods and based on the landmark model methods and so on.