Search

Recent years have witnessed the rapid development of the China's stock market, but investment risks have also emerged. Stock price is always unstable and non-linear, affected not only by historical transaction data but also by national policies, news, and other data. Stock price and textual data are beginning to be employed in the prediction process. However, the challenge lies in effectively integrating feature information derived from stock price and textual information. To address the problem, in this paper, this paper proposes a Hybrid Data-driven Multi-task Learning(HDML) framework to predict stock price. HDML adopts hybrid data as model input, mining the transaction and capital flow data information in the stock market and considering the impact of investors' emotions on the stock market. In addition, we incorporate multi-task learning, which predicts the closing price range of stock based on structured data and then corrects the prediction results through investors' comment text data. HDML effectively captures the relationship between different modal data through multi-task learning and achieve improvements on both tasks. The experimental results show that compared with previous work, HDML reduces the RMSE of the evaluation set by 12.14% and improves the F1 score by an average of 13.64% at the same time. Moreover, value at risk (VaR), together with the HDML model, can help investors weigh the potential gains against the associated risks. [ABSTRACT FROM AUTHOR]

With the development of the big data era, the amount of data has entered an explosive growth phase. Limited by the constraints of cost, efficiency, and security of self-built storage systems, enterprises are forced to outsource files to cloud storage systems. However, the lack of file security and auditability in cloud storage systems continues to threaten the security of outsourced files. This paper designs and implements a large file security outsourcing and sharing method based on blockchain (BFSOut). It uses Hyperledger Fabric and Interplanetary File System (IPFS) as the underlying storage engine, which solves the problem of cloud storage security issues. In order to ensure the security of outsource files, the client-side offline encryption and decryption is used. Furthermore, a dynamic hybrid encryption scheme is adopted, making the overall encryption effect more efficient. Meanwhile, BFSOut effectively reduces the time consumption of large file encryption and decryption by 20%–30% through the parallel block encryption and decryption model. Experimental performance and security analysis show that the method has good performance and safety. [ABSTRACT FROM AUTHOR]

Automatic summarization is attracting increasing attention as one of the most promising research areas. This technology has been tried in various real-world applications in recent years and achieved a good response. However, the applicability of conventional evaluation metrics cannot keep up with rapidly evolving summarization task formats and ensuing indicator. After recent years of research, automatic summarization task requires not only readability and fluency, but also informativeness and consistency. Diversified application scenarios also bring new challenges both for generative language models and evaluation metrics. In this review, we analysis and specifically focus on the difference between the task format and the evaluation metrics. [ABSTRACT FROM AUTHOR]

With the continuous and rapid development of cloud-based data services, the total volume of Internet data is experiencing explosive growth. Nevertheless, contemporary centralized cloud storage-oriented data service providers encounter significant challenges in fully satisfying the requirements for user data security, fine-grained access control, and consistently high-performance standards. In this paper, we propose and implement an efficient peer-to-peer data storage and sharing system to address the challenges faced by service providers. Our solution utilizes the modified EOSIO blockchain and the InterPlanetary File System (IPFS) distributed storage as the underlying data storage framework. We employ hybrid encryption to ensure the secrecy of the shared data of the users and also to facilitate multiple uses and persistent storage of the shared data of the users. Moreover, with the expansion of the blockchain component, we provide a flexible transaction information audit solution that helps to trace the source of malicious behavior and reduces the cost of using blockchain information. In system analysis and experimental evaluation, compared to traditional blockchain storage, FIBPRO has theoretically achieved a 98.76% reduction in on-chain storage consumption. In practical concurrency testing, it achieved a comprehensive performance of approximately 1300 TPS (transactions per second), with an average upload efficiency of about 2.31MB/s and a download rate of about 5.29MB/s. These results demonstrate the system's availability and scalability. [ABSTRACT FROM AUTHOR]

Many real-world applications require prediction that takes the most advantage of data. Classic data mining mechanisms tend to feed a prediction model pivotal data to achieve a promising result, which needs to be adjusted in different application scenarios. Recent studies have shown the potential of I Ching mechanism to improve prediction capacity. However, the I Ching prediction mechanism has several issues, including underutilized I Ching knowledge and incomplete data conversion. To address these issues, the authors designed a model to leverage I Ching knowledge and transform traditional I Ching prediction processing into data mining. The authors' investigation revealed promising results in the stock market compared to popular machine learning and deep learning algorithms such as support vector machine (SVM), extreme gradient boosting (XGBoost), and long short-term memory (LSTM). [ABSTRACT FROM AUTHOR]

With the development of China's economy, the urban floating population is also increasing, resulting in a sharp increase in the amount of urban waste. How to recycle and dispose of municipal waste more efficiently has become the top concern of municipalities and other relevant departments. In this article, the above problem is transformed into the municipal waste collection vehicle routing problem (MWCVRP) to solve the problem with the minimum total waste transportation cost. Because the carrying capacity of different models is different, this article introduces a cost calculation criterion that combines the total mileage of different models of transport vehicles and the number of station services. A multi-model garbage truck path optimization model is established, and then a heuristic-based task dynamic assignment algorithm is designed to solve the problem. The Solomon dataset is used to verify the feasibility and effectiveness of the model and algorithm through experiments. [ABSTRACT FROM AUTHOR]

System logs are almost the only data that records system operation information, so they play an important role in anomaly analysis, intrusion detection, and situational awareness. However, it is still a challenge to obtain effective data from massive system logs. On the one hand, system logs are unstructured data, and, on the other hand, system log records cannot be directly analyzed and calculated by computers. In order to deal with these problems, current researchers digitize system logs through two key steps of log parsing and feature extraction. This paper classifies, analyzes, and summarizes the current log analysis research in terms of log parsing and feature extraction by investigating articles in recent years (including ICSE, TKDD, ICDE, IJCAI, ISSRE, ICDM, ICWS, ICSME, etc.). Finally, in combination with the existing research, the research prospects in the field are elaborated and predicted. [ABSTRACT FROM AUTHOR]

As the most successful cryptocurrency, bitcoin has become the primary target of attackers. The security risks existing in bitcoin network (P2P networks) may pose serious threats to itself. It has been proved that network attackers of the autonomous system level could isolate a specific set of bitcoin nodes using prefix hijacking attacks; since this attack achieves bitcoin partition by deleting all data packets of the victim node, it is easy to be discovered by the victim node, and cannot maintain a long-term connection (the partition will disappear after canceling the BGP hijacking) (Apostolaki M et al. (2017)). This paper proposes a new attack scheme—eclipse attack method based on BGP hijacking (BHE). The attack can occupy the network connection of the victim node, and only need to delete part of the TCP handshaking packets of the victim node during the attack, and it makes the attack more hidden and can occupy the network connection of the victim node for a long time. The innovation of the BHE attack is that it can control the peering decision of the victim node by controlling the victim node's internal peer database (new table and tried table) and preventing the victim node from establishing a good connection. It enables the attacker to occupy all network connections of the victim node and become its natural network middleman. We verify the feasibility of the BHE attack through experimental evaluation and demonstrate that an attacker who can launch BGP hijacking may occupy all connections of the victim node within 20 minutes (ignoring the time of traffic diversion). To reduce the attack's impact, the paper provides some countermeasures that can use in practice according to the basic characteristics of the attack. [ABSTRACT FROM AUTHOR]

Recently, machine learning techniques, especially supervised learning techniques, have been adopted in the Intrusion Detection System (IDS). Due to the limit of supervised learning, most state-of-the-art IDSs do not perform well on unknown attacks and incur high computational overhead in the Internet of Things (IoT). To overcome these challenges, we propose a novel IDS based on unsupervised techniques, namely, UTEN-IDS. UTEN-IDS uses the ensemble of autoencoders to handle the network data and performs the anomaly detection by an Isolation Forest algorithm. The effectiveness of the proposed method is verified using two benchmark datasets. The results show that our approach has significant advantages in classification performance and proves its utility in the IoT network when compared to other approaches. [ABSTRACT FROM AUTHOR]

Public chains represented by Bitcoin and Ethereum do not require users to use their real names, and transaction data are open to the whole network. Analysed based on this, researchers have achieved the deanonymization of blockchain transactions to a certain extent. Based on the existing blockchain transaction privacy protection scheme, the true link relationship between the transaction sender and receiver is hidden, which brings difficulties to regulation. In this paper, we propose a cryptocurrency mixing service RBSmix, which allows users to reestablish their financial privacy in Bitcoin and related cryptocurrencies. RBSmix, through blind signature to prevent attackers from linking input and output addresses, by the threshold secret sharing algorithm, encryption technology, and a regulation team, combined with the idea of voting, tracks the source of funds for illegal addresses. Experiments show that the scheme scales to large numbers of users and can provide users with better privacy protection. [ABSTRACT FROM AUTHOR]

Neural networks have been proved to perform well in network intrusion detection. In order to acquire better features of network traffic, more learning layers are necessarily required. However, according to the results of the previous research, adding layers to the neural networks might fail to improve the classification results. In fact, after the number of layers has reached a certain threshold, performance of the model tends to degrade. In this paper, we propose a network intrusion detection model based on residual learning. After transforming the UNSW-NB15 data set into images, deeper convolutional neural networks with residual blocks are built to learn more critical features. Instead of the cross-entropy loss function, the modified focal loss is calculated to address the class imbalance problem in the training set and identify minor attacks in the testing set. Batch normalization and global average pooling are used to avoid overfitting and enhance the model. Experimental results show that the proposed model can improve attack detection accuracy compared with existing models. [ABSTRACT FROM AUTHOR]

With the rapid popularity of social network applications and advanced digital devices, we have witnessed the explosive growth of multimedia big data in terms of both scale and variety over the last few years. Yet at the same time, security issues related to multimedia big data have arisen. An urgent demand has emerged for novel technologies that deal with copyright protection, multimedia forgery detection, and cybersecurity—especially for cyber-enabled multimedia applications. This special issue brings together research efforts in cybersecurity for cyber-enabled multimedia applications. [ABSTRACT FROM AUTHOR]

Recovery of fragmented files is an important part of digital forensics. Video files are more likely to be fragmented since their sizes are relatively large that recovering video files without the file system information is meaningful. This paper presents a video recovery technique of a fragmented video file using the frame size information in every frame and the index. Many existing video recovery techniques attempt to recover videos using file system information or header/footer flag. These approaches may fail in the situations which the file system information is unknown or videos are fragmented. The proposed method addresses how to extract AVI file fragments from data images and map all the extracted fragments into original order. Experiments result show that mapping the AVI fragments according to the frame size information in every fragment and index is credible and the non-overwritten part of the fragmented video can be recovered using the method. [ABSTRACT FROM AUTHOR]