Search

In forensic document analysis, the authenticity of a document must be properly checked in the context of suspected forgery. Hyperspectral Imaging (HSI) is a non-invasive way of detecting fraudulent papers in a multipage document. The occurrence of a forged paper in a multi-page document may have a substantial difference from rest of the papers in its age, type, color, texture, and so on. Each pixel in an HSI data can be used as the material fingerprint for the spatial point it corresponds to. Hence, hyperspectral data of paper samples made of the same substance have similar characteristics and can be grouped into a single cluster. Similarly, paper samples made of different substances have different spectral properties. This paper relies on this heuristic and proposes a tensor based clustering framework for hyperspectral paper data, with an application to detect the forged papers in multi-page documents. Information embedded in the hyperspectral patches of the papers to be clustered is arranged into individual lateral slices of a third-order tensor in this framework. Further, this work employs the self-expressiveness property of submodules and an objective function is formulated to extract self-expressive representation tensor with low multirank and f-diagonal structure. Objective function of the proposed method incorporates $l_{\frac {1}{2}}$ -induced Tensor Nuclear Norm (TNN) and $l_{\frac {1}{2}}$ regularization to impart better low rankness and f-diagonal structure to the representation tensor. Experimental results of the proposed method were compared to the state-of-the-art subspace clustering approaches. The results demonstrate improved performance of the proposed method over the existing clustering algorithms.

Introduction: Due to the high prevalence of depression and the considerable pressure it puts on individuals, society and healthcare system, it is essential to conduct sufficient research to help with decision making in prevention, treatment and control of this condition. To assist with research planning and setting priorities, this paper aimed to identify the thematic structure of papers related to depression treatment. Methods: This was an applied research study which used Scientometrics approach. To obtain data, the keyword “Depression” was searched as a mesh descriptor with the subdivision “Therapy” in PubMed database for the period of 2005 to 2014. The data were analyzed using co-word and clustering methods with the help of Excel and SPSS software packages. Results: The growth of scientific production in this field appeared to have been balanced throughout the ten-year period examined in the study. The most active research areas for the two five year periods (2005-2009 and 2010-2014) were drug therapy, psychology, and medication adherence. Two hierarchical graphs of descriptors for each five-year period were prepared both of which were composed of 12 clusters with 34 common descriptors. Conclusion: The findings based on the inclusion index showed that only 20 percent of topics in the second five-year period of the study were novel. Therefore, it could be concluded that research areas that were related with each other in previous years will probably co-occur with other topics in the future might disappear. In every discipline, some topics are considered to be fundamental and research works are being carried out on them almost every year.

Renewable energy production throughout low-voltage grids has gradually increased in electrical distribution systems, therefore introducing small energy producers - prosumers. This paradigm challenges the traditional unidirectional energy distribution flow to include disperse power production from renewables. To understand how energy usage can be optimized in the dynamic electrical grid, it is important to understand the behavior of prosumers and their impact on the grid's operational procedures. The main focus of this study is to investigate how grid operators can obtain an automatic data-driven system for the low-voltage electrical grid management, by analyzing the available grid topology and time-series consumption data from a real-life test area. The aim is to argue for how different consumer profiles, clustering and prediction methods contribute to the grid-related operations. Ultimately, this work is intended for future research directions that can contribute to improving the trade-off between systematic and scalable data models and software computational challenges. [ABSTRACT FROM AUTHOR]

Gene expression responses of paper birch (Betula papyrifera) leaves to elevated concentrations of CO2 and O3 were studied with microarray analyses from three time points during the summer of 2004 at Aspen FACE. Microarray data were analyzed with clustering techniques, self-organizing maps, K-means clustering and Sammon''s mappings, to detect similar gene expression patterns within sampling times and treatments. Most of the alterations in gene expression were caused by O3, alone or in combination with CO2. O3 induced defensive reactions to oxidative stress and earlier leaf senescence, seen as decreased expression of photosynthesis- and carbon fixation-related genes, and increased expression of senescence-associated genes. The effects of elevated CO2 reflected surplus of carbon that was directed to synthesis of secondary compounds. The combined CO2 +O3 treatment resulted in differential gene expression than with individual gas treatments or in changes similar to O3 treatment, indicating that CO2 cannot totally alleviate the harmful effects of O3. [Copyright &y& Elsevier]

This paper overviews an assortment of recent research work undertaken on recommender system models based on using multiple views of user and item-related data across the recommendation process. A summary of representative literature on multi-view recommender approaches is provided, describing their main characteristics, such as: their potential to overcome most common shortcomings in conventional recommender systems, as well as the use of data science, learning techniques and aggregation processes to combine information stemming from multiple views. A tabular summary is provided to facilitate the comparison of the similarities and differences among the surveyed works, along with commonly identified directions for future research in the topic. [ABSTRACT FROM AUTHOR]

Copyright of Problems of Economy is the property of Research Centre for Industrial Developmen Problems of Nas and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)

Data nowadays is the language of technologies as every process needs a data to be processed the input is data and the output also is data. Analyzing the data is a significant task especially with the increasing production of the data particularly data as a text, it would be difficult to manually analyze the data, extract information and detect the hidden patterns from unstructured text. Datamining is automated technique for gathering or deriving a new high-quality information and uncover the relations among the data. Text mining is one of main branches of the data mining however data mining this paper, an is more comprehensive overview for mining the publication papers via text mining techniques and their results and evaluation would be presentedas following: the first approachis keywords extraction using natural language processing (NLP) approach, the second approach named entity recognition and the last approach is document clustering where machine learning techniques are applied to the both of them. [ABSTRACT FROM AUTHOR]

Food loss ans waste, referred next as FLW, has a great extent on the economic and behavioural patterns of consumption. In this paper, we aim to present statistical univariate and multivariate analyses based on a statistical study run during a period of several months in the Romanian territory. In this matter, the main purpose of the analysis developed in this paper is to determine several patterns of FLW phenomenon. This is made based on a methodology which comprises two main sections: the univariate analysis, consisting in statistical determinations of indicators and direct observation, and the multivariate analysis, related to the clustering analysis. The clustering analysis is based on the usage of hierarchical method, which has as final result the determination of clusters describing patterns of consumption related to food loss and waste. After the analysis was run, five main clusters related to food waste behaviour patterns were determined, with slight differences between food consumption habits, but with various combinations between the parameters taken into consideration. The obtained clusters offer important information for policymakers and stakeholders aiming to customize interventions and programs to target specific demographic groups or segments of the population. [ABSTRACT FROM AUTHOR]

By the end of 2020, over 16,300 Massive Open Online Courses (MOOCs) from 950 universities worldwide had enrolled over 180 million students. Interest in MOOCs has been matched by significant research on the topic, including a considerable number of reviews. This study uses Machine Learning techniques and human expert supervision to generate a comprehensive systematic literature mapping review that overcomes some limitations of the traditional ones and provides a broader overview of the content and main topics studied in the specialized literature devoted to MOOCs. The sample consisted of 6320 publications automatically classified within six research topics, denominated by human experts: institutional approach, pedagogical approach, evaluation, analytics, participation, and educational resources. The content analysis of the topics identified was conducted using visual network analysis, which supported the identification of different thematic sub-clusters and endorsed the classification. Results from the review show that the lowest production of MOOC papers is within the topics of the pedagogical approach and educational resources. In contrast, participation and evaluation are the most frequent ones. In addition, the most cited papers are on the topics of analytics and resources, being the pedagogical approach and the institutional approach the less cited. This highlights the need for more MOOC research from a pedagogical perspective and calls upon the presence of educators. [ABSTRACT FROM AUTHOR]

The Internet of Things (IoT) is a very famous network because of its many applications. IoT network has an integration of large-scale IoT devices that generate data. These IoT devices are very low power computing devices due to which they have a low level of communication. These devices construct data and transmit the data to the base station via intermediate IoT devices. The base station gathers and integrates the data and sends it to the administrator for further processing. The data attains the base station using various routing algorithms with the goal of low power consumption. When discussing low power IoT devices, power efficiency is an important performance measurement when creating a routing algorithm. This paper proposes a Minimum Power Consumption Routing (MPCR) algorithm using Hierarchical Fuzzy Logic Clustering (HFLC) algorithm for IoT networks. The MPCR with HFLC algorithm is an energy-efficient algorithm because of its lower power consumption for the cluster by aggregating the data within the cluster head and decreasing the number of data transmissions to the base station. In this paper, the cluster formation and cluster-head selection are explained, and a simulation has been conducted. In addition, the proposed algorithm is compared with the existing algorithms based on different metrics such as throughput, packet delivery ratio, and energy consumption of the network. The experimental results show that the proposed MPCR with the HFLC algorithm provides high throughput and packet delivery ratio and reduces energy consumption more efficiently than other existing algorithms. [ABSTRACT FROM AUTHOR]

New discoveries and knowledge are summarized in thousands of published papers per year per scientific domain, making it incomprehensible for scientists to account for all available knowledge relevant for their studies. In this paper, we present ACCORDION (ACCelerating and Optimizing model RecommenDatIONs), a novel methodology and an expert system that retrieves and selects relevant knowledge from literature and databases to recommend models with correct structure and accurate behavior, enabling mechanistic explanations and predictions, and advancing understanding. ACCORDION introduces an approach that integrates knowledge retrieval, graph algorithms, clustering, simulation, and formal analysis. Here, we focus on biological systems, although the proposed methodology is applicable in other domains. We used ACCORDION in nine benchmark case studies and compared its performance with other previously published tools. We show that ACCORDION is: comprehensive, retrieving relevant knowledge from a range of literature sources through machine reading engines; very effective, reducing the error of the initial baseline model by more than 80%, recommending models that closely recapitulate desired behavior, and outperforming previously published tools; selective, recommending only the most relevant, contextspecific, and useful subset (15%-20%) of candidate knowledge in literature; diverse, accounting for several distinct criteria to recommend more than one solution, thus enabling alternative explanations or intervention directions. [ABSTRACT FROM AUTHOR]

With the rapid development of internet communication and the wide application of intelligent terminal, moving the cache to the edge of the network is an effective solution to shorten the delay of users accessing content. However, the existing cache work lacks the comprehensive consideration of users and content, resulting in low cache hit ratio and low accuracy of the whole system. In this paper, the authors propose a collaborative caching model that considers both user request content and content prediction, so as to improve the caching performance of the whole network. Firstly, the model uses the clustering algorithm based on Akike information criterion to cluster users. Then, combined with the clustering results, echo state network is used as the machine learning framework to predict the content. Finally, the cache contents are selected according to the prediction results and cached in the cache unit of the small base station. Simulation results show that compared with the existing cache algorithms, the proposed method has obvious improvement in cache hit ratio, accuracy, and recall rate. [ABSTRACT FROM AUTHOR]

Reverse mortgage is one of the products (perhaps the main one) that is good to obtain additional income by using the habitual residence as collateral. The main objective of this paper is to analyse the effects that reverse mortgage contracting has on household finances over the lifetime of a family according to the socioeconomic group to which it belongs in Spain. Four indicators are employed to measure the immediate and long-term effects. We use a stochastic model with a double source of randomness, survival and entry into dependency, and apply it to the three socioeconomic groups obtained with cluster methodology from the 2017 Spanish Household Financial Survey data. We conclude that the effects are very different depending on the group: regarding only the effects of hiring a reverse mortgage on the income of the family, widowed women aged between 81 and 85 years, with low income and expenses as well as little net wealth, and a habitual residence that represents half of her net wealth (Cluster 1) are the most benefited; considering that the highest impact indicators are on the probability of illiquidity and on the value of lack of liquidity, the use of reverse mortgages benefits more the families in Cluster 3 (high income and expenses and really high net wealth, head of household aged between 76 and 80 years) and less the families in Cluster 2 (medium income, net wealth and expenses, head of household aged between 65 and 75 years). [ABSTRACT FROM AUTHOR]

Research productivity and impact (RPI) is commonly measured through citation analysis, such as the h-index. Despite the popularity and objectivity of this type of method, it is still difficult to effectively compare a number of related researchers in terms of various citation-related statistics at the same time, such as average cites per year/paper, the number of papers/citations, h-index, etc. In this work, we develop a method that employs information visualization technology, and examine its applicability for the assessment of researchers' RPI. Specifically, our prototype, a visualizing research productivity and impact (VisualRPI) system, is introduced, which is composed of clustering and visualization components. The clustering component hierarchically clusters similar research statistics into the same groups, and the visualization component is used to display the RPI in a clear manner. A case example using information for 85 information systems researchers is used to demonstrate the usefulness of VisualRPI. The results show that this method easily measures the RPI for various performance indicators, such as cites/paper and h-index. [ABSTRACT FROM AUTHOR]

Traditional metrics may not adequately assess performance in certain situations, whereas the Area Under Curve (AUC) offers a comprehensive perspective by considering both sensitivity and specificity. This method enhances interpretability, addresses limitations, and promotes the development of robust clustering algorithms. In unsupervised learning, utilizing AUC is a significant method for improving the precision and accuracy of machine learning models. Our work is inspired by several recent related works that implement approaches to manage the challenges of developing new metrics that can effectively assess and evaluate the performance of clustering algorithms. The research question relies on the concept of using an optimal metric for model evaluation of classification and clustering. Therefore, the paper investigates the use of the classification metric AUC for clustering validation purposes. The methodology we adopt is a hybrid clustering model because such a technique offers a robust model by combining the strengths of each model. The linkage approach directly impacts the clustering results, so we give significant attention to this feature in our implementation. Among the various linkage methods, we utilized single and average linkages. The Manhattan and Euclidean metrics are the distance measures used in this work. Thus, our contribution is to explore the benefit of using linkages and distance measurement in clustering with the help of the AUC metric. In addition, the entire proposed work and the contributions of this paper are evaluated and applied to the NSL-KDD dataset. Based on the proposed approach of using AUC with clustering, the Detection Rate (DR), False Alarm Rate (FAR), and other criteria are chosen to examine the model's results and capabilities. [ABSTRACT FROM AUTHOR]

The problem of image-based detection of the variety of beans, using artificial intelligence, is currently dealt with by scientists of various profiles. The idea of this paper is to show the possibility of applying different types of distances, primarily those that are fuzzy metrics, in clustering models in order to improve existing models and obtain more accurate results. The paper presents the method of variable neighborhood search, which uses both standard and fuzzy t-metrics and dual fuzzy s-metrics characterized by appropriate parameters. By varying those parameters of the fuzzy metric as well as the parameters of the metaheuristic used, we have shown how it is possible to improve the clustering results. The obtained results were compared with existing ones from the literature. The criterion function used in clustering is a fuzzy metric, which is proven in the paper. [ABSTRACT FROM AUTHOR]

Current edge devices for neural networks such as FPGA, CPLD, and ASIC can support low bit-width computing to improve the execution latency and energy efficiency, but traditional linear quantization can only maintain the inference accuracy of neural networks at a bit-width above 6 bits. Different from previous studies that address this problem by clipping the outliers, this paper proposes a two-stage quantization method. Before converting the weights into fixed-point numbers, this paper first prunes the network by unstructured pruning and then uses the K-means algorithm to cluster the weights in advance to protect the distribution of the weights. To solve the instability problem of the K-means results, the PSO (particle swarm optimization) algorithm is exploited to obtain the initial cluster centroids. The experimental results on baseline deep networks such as ResNet-50, Inception-v3, and DenseNet-121 show the proposed optimized quantization method can generate a 5-bit network with an accuracy loss of less than 5% and a 4-bit network with only 10% accuracy loss as compared to 8-bit quantization. By quantization and pruning, this method reduces the model bit-width from 32 to 4 and the number of neurons by 80%. Additionally, it can be easily integrated into frameworks such as TensorRt and TensorFlow-Lite for low bit-width network quantization. [ABSTRACT FROM AUTHOR]

Most cybersecurity frameworks are based on three major components such as confidentiality, integrity, and availability. All these components have their parameters that are used to secure network nodes. But finding the most cyber secure node in a network needs a measurement method. The aim of the paper is to offer a model that can be used to find the most secure network nodes considering these cybersecurity components and their parameters. This is achieved by modelling numeric values of respective weights for parameters of confidentiality, integrity, and availability. The model is applied to a simulated environment where random values standing for cybersecurity parameters are given to 30 wireless network nodes that are used as an example. Then the weighted values are processed with Python programming language by giving the most secure nodes according to needed cybersecurity components. This model can be used to recommend the right network node that can be used to deploy services securely while avoiding potential vulnerabilities and cyber-attacks. [ABSTRACT FROM AUTHOR]

Enhancing the management and monitoring of oil and gas processes demands the development of precise predictive analytic techniques. Over the past two years, oil and its prediction have advanced significantly using conventional and modern machine learning techniques. Several review articles detail the developments in predictive maintenance and the technical and non-technical aspects of influencing the uptake of big data. The absence of references for machine learning techniques impacts the effective optimization of predictive analytics in the oil and gas sectors. This review paper offers readers thorough information on the latest machine learning methods utilized in this industry's predictive analytical modeling. This review covers different forms of machine learning techniques used in predictive analytical modeling from 2021 to 2023 (91 articles). It provides an overview of the details of the papers that were reviewed, describing the model's categories, the data's temporality, field, and name, the dataset's type, predictive analytics (classification, clustering, or prediction), the models' input and output parameters, the performance metrics, the optimal model, and the model's benefits and drawbacks. In addition, suggestions for future research directions to provide insights into the potential applications of the associated knowledge. This review can serve as a guide to enhance the effectiveness of predictive analytics models in the oil and gas industries. [ABSTRACT FROM AUTHOR]

The aim of this paper is to examine the impact of a country's entry into the monetary union on its economic variables growth rate of real GDP as well as on GDP per capita growth for the period from 2010 to 2020. The clustering method and the ANFIS method were used in the data analysis. A total of two cluster analyses were performed. The first cluster includes countries that joined the EU in 2004 and became EZ members by 2010. The second cluster refers to those countries that joined the EU in 2004 but are not yet members of the EZ. For the first individual cluster analysis two models were analysed and for the second individual cluster three models were analysed using the ANFIS method. As expected, the results showed that GDP growth is connected with trade, inflation and gross investments in fixed capital in the observed countries, while GDP per capita is connected with unemployment, interest rates and public debt. With regard to GDP growth, the difference between countries that are in the eurozone and those that are not is not significant, which is in line with other studies. [ABSTRACT FROM AUTHOR]

Superpixel segmentation can get the middle features in image processing, effectively reduce the dimensionality of the image, and is widely used in image processing fields. To get the regular and compact superpixels in real-time, a superpixel segmentation algorithm based on image density is proposed in this paper. Firstly, the image is uniformly divided according to the number of superpixels to be obtained. Secondly, to get the clustering ability of the pixels, the density image is produced. Thirdly, the seed is chosen in each sub-region block according to the density and then the superpixels are obtained by clustering. During the clustering process, the pixel around the seed should be added into the superpixel if it meets the conditions, and the small supeipixels are merged into the big superpixels around them. Finally, the result shows that the proposed algorithm has the best segmentation effect, and a good balance in accuracy, regularity, and time cost. [ABSTRACT FROM AUTHOR]

An important part of diagnostics is to gain insight into properties that characterize a disease. Machine learning has been used for this purpose, for instance, to identify biomarkers in genomics. However, when patient data are presented as images, identifying properties that characterize a disease becomes far more challenging. A common strategy involves extracting features from the images and analyzing their occurrence in healthy versus pathological images. A limitation of this approach is that the ability to gain new insights into the disease from the data is constrained by the information in the extracted features. Typically, these features are manually extracted by humans, which further limits the potential for new insights. To overcome these limitations, in this paper, we propose a novel framework that provides insights into diseases without relying on handcrafted features or human intervention. Our framework is based on deep learning (DL), explainable artificial intelligence (XAI), and clustering. DL is employed to learn deep patterns, enabling efficient differentiation between healthy and pathological images. Explainable artificial intelligence (XAI) visualizes these patterns, and a novel "explanation-weighted" clustering technique is introduced to gain an overview of these patterns across multiple patients. We applied the method to images from the gastrointestinal tract. In addition to real healthy images and real images of polyps, some of the images had synthetic shapes added to represent other types of pathologies than polyps. The results show that our proposed method was capable of organizing the images based on the reasons they were diagnosed as pathological, achieving high cluster quality and a rand index close to or equal to one. [ABSTRACT FROM AUTHOR]

This paper proposes a method to reduce seismogram variability as a determining factor in its interpretation, processing, and clustering. By introducing the concept of single fluctuations (SFs), the seismogram can be parsed into a sequence of random impulses with subsequent ordering. This rearrangement of SFs, if they are assigned by positive integers, represents the formal mapping of a regular string of integers into a random one, which can be represented with a bipartite random graph (bigraph). Due to its specific randomness, such a bigraph is considered a new type of random balanced bigraph. The R-envelope and RQ-envelope, its equidistant version, are defined by tracing the peak envelope over ordered SFs. The equivalence and complementariness of the RQ and bigraph are considered and discussed, forming a combined characteristic of the seismogram. The R/RQ provided a considerable reduction in seismogram variability, which was confirmed by creating and analyzing an ensemble of RQ from several seismograms. In the RQ domain, distance is defined as a possible basis for metrics and clustering, but the ensemble variability was quite narrow and not as suitable for this purpose. Otherwise, the ensemble shows high redundancy hidden in the seismogram population. As for the bigraph, the mesh of its edges is structuralized in bundles, forming a skeleton, which reflects the internal structural content of the seismogram. The distance over the domain of bigraphs is proposed to show the possibility of clustering. This means that only a combined RQ and bigraph provides a suitable frame for seismogram representation with reduced variability and, thus, the potential for more effectively organizing seismic databases and a deeper interpretation of seismograms; therefore, RQs and bigraphs can be considered as a transform of a seismogram. Many aspects of these concepts are thoroughly discussed. The similarity between concepts of SF and wavelets is briefly considered. This very complex theme is new and promises broad further research. All issues considered in the paper are abundantly illustrated. [ABSTRACT FROM AUTHOR]