Semantic Features with Contextual Knowledge-Based Web Page Categorization Using the GloVe Model and Stacked BiLSTM

A prospective approach to detect advanced persistent threats: Utilizing hybrid optimization technique

Advanced Persistent Threat (APT) attacks pose significant challenges for AI models in detecting and mitigating sophisticated and highly effective cyber threats. This research introduces a novel concept called Hybrid HHOSSA which is the grouping of Harris Hawk Optimization (HHO) and Sparrow Search Algorithm (SSA) characteristics for optimizing the feature selection and data balancing in the context of APT detection. In addition, the light GBM as well as the weighted average Bi-LSTM are optimized by the proposed hybrid HHOSSA optimization. The HHOSSA-based attribute selection is used to choose the most important attributes from the provided dataset in the early step of the quasi-identifier detection. The HHOSSA-SMOTE algorithm effectively balances the unbalanced data, such as the lateral movements and the data exfiltration in the DAPT 2020 database, which further improves the classifier performance. The light GBM and the Bi-LSTM classifier hyperparameters are well attuned and classified by the HHOSSA optimization for the precise classification of the attacks. The outcome of both the optimized light GBM and the Bi-LSTM classifier generates the final prediction of the attacks existing in the network. According to the research findings, the HHOSSA-hybrid classifier achieves high accuracy in detecting attacks, with an accuracy rate of 94.468 %, a sensitivity of 94.650 %, and a specificity of 95.230 % with a K-fold value of 10. Also, the HHOSSA-hybrid classifier achieves the highest AUC percentage of 97.032, highlighting its exceptional performance in detecting APT attacks.

Short text classification with machine learning in the social sciences: The case of climate change on Twitter

To analyse large numbers of texts, social science researchers are increasingly confronting the challenge of text classification. When manual labeling is not possible and researchers have to find automatized ways to classify texts, computer science provides a useful toolbox of machine-learning methods whose performance remains understudied in the social sciences. In this article, we compare the performance of the most widely used text classifiers by applying them to a typical research scenario in social science research: a relatively small labeled dataset with infrequent occurrence of categories of interest, which is a part of a large unlabeled dataset. As an example case, we look at Twitter communication regarding climate change, a topic of increasing scholarly interest in interdisciplinary social science research. Using a novel dataset including 5,750 tweets from various international organizations regarding the highly ambiguous concept of climate change, we evaluate the performance of methods in automatically classifying tweets based on whether they are about climate change or not. In this context, we highlight two main findings. First, supervised machine-learning methods perform better than state-of-the-art lexicons, in particular as class balance increases. Second, traditional machine-learning methods, such as logistic regression and random forest, perform similarly to sophisticated deep-learning methods, whilst requiring much less training time and computational resources. The results have important implications for the analysis of short texts in social science research.

Contextual Embeddings-Based Web Page Categorization Using the Fine-Tune BERT Model

The World Wide Web has revolutionized the way we live, causing the number of web pages to increase exponentially. The web provides access to a tremendous amount of information, so it is difficult for internet users to locate accurate and useful information on the web. In order to categorize pages accurately based on the queries of users, methods of categorizing web pages need to be developed. The text content of web pages plays a significant role in the categorization of web pages. If a word’s position is altered within a sentence, causing a change in the interpretation of that sentence, this phenomenon is called polysemy. In web page categorization, the polysemy property causes ambiguity and is referred to as the polysemy problem. This paper proposes a fine-tuned model to solve the polysemy problem, using contextual embeddings created by the symmetry multi-head encoder layer of the Bidirectional Encoder Representations from Transformers (BERT). The effectiveness of the proposed model was evaluated by using the benchmark datasets for web page categorization, i.e., WebKB and DMOZ. Furthermore, the experiment series also fine-tuned the proposed model’s hyperparameters to achieve 96.00% and 84.00% F1-Scores, respectively, demonstrating the proposed model’s importance compared to baseline approaches based on machine learning and deep learning.

Predicting User Satisfaction of Mobile Healthcare Services using Machine Learning

Outbreak of the COVID-19 leads to rapid development of the mobile healthcare services. Given that user satisfaction is of great significance in inducing marketing success in competition markets, this research explores and predicts user satisfaction with mobile healthcare services. Specifically, the current research aimed to design a machine learning model that predicts user satisfaction with healthcare services using big data from Google Play Store reviews and satisfaction ratings. By dealing with the sentimental features in online reviews with five classifiers, the authors find that Logistic regression with term frequency-inverse document frequency (TF-IDF) and XGBoost with Bag of words (BoW) have superior performances in predicting user satisfaction for healthcare services. Based on these results, the authors conclude that such user-generated texts as online reviews can be used to predict user satisfaction, and Logistic regression with TF-IDF and XGBoost with BoW can be prioritized for developing online review analysis platforms for healthcare service providers.

A microblog content credibility evaluation model based on collaborative key points

The spread of false content on microblogging platforms has created information security threats for users and platforms alike. The confusion caused by false content complicates feature selection during credibility evaluation. To solve this problem, a collaborative key point-based content credibility evaluation model, CECKP, is proposed in this paper. The model obtains the key points of the microblog text from the word level to the sentence level, then evaluates the credibility according to the semantics of the key points. In addition, a rumor lexicon constructed collaboratively during word-level coding strengthens the semantics of related words and solves the feature selection problem when using deep learning methods for content credibility evaluation. Experimental results show that, compared with the Att-BiLSTM model, the F1 score of the proposed model increases by 3.83% and 3.8% when the evaluation results are true and false respectively. The proposed model accordingly improves the performance of content credibility evaluation based on optimized feature selection.

A Deep Learning Framework for Detection of COVID-19 Fake News on Social Media Platforms

The fast growth of technology in online communication and social media platforms alleviated numerous difficulties during the COVID-19 epidemic. However, it was utilized to propagate falsehoods and misleading information about the disease and the vaccination. In this study, we investigate the ability of deep neural networks, namely, Long Short-Term Memory (LSTM), Bi-directional LSTM, Convolutional Neural Network (CNN), and a hybrid of CNN and LSTM networks, to automatically classify and identify fake news content related to the COVID-19 pandemic posted on social media platforms. These deep neural networks have been trained and tested using the “COVID-19 Fake News” dataset, which contains 21,379 real and fake news instances for the COVID-19 pandemic and its vaccines. The real news data were collected from independent and internationally reliable institutions on the web, such as the World Health Organization (WHO), the International Committee of the Red Cross (ICRC), the United Nations (UN), the United Nations Children’s Fund (UNICEF), and their official accounts on Twitter. The fake news data were collected from different fact-checking websites (such as Snopes, PolitiFact, and FactCheck). The evaluation results showed that the CNN model outperforms the other deep neural networks with the best accuracy of 94.2%.

LogLS: Research on System Log Anomaly Detection Method Based on Dual LSTM

System logs record the status and important events of the system at different time periods. They are important resources for administrators to understand and manage the system. Detecting anomalies in logs is critical to identifying system faults in time. However, with the increasing size and complexity of today’s software systems, the number of logs has exploded. In many cases, the traditional manual log-checking method becomes impractical and time-consuming. On the other hand, existing automatic log anomaly detection methods are error-prone and often use indices or log templates. In this work, we propose LogLS, a system log anomaly detection method based on dual long short-term memory (LSTM) with symmetric structure, which regarded the system log as a natural-language sequence and modeled the log according to the preorder relationship and postorder relationship. LogLS is optimized based on the DeepLog method to solve the problem of poor prediction performance of LSTM on long sequences. By providing a feedback mechanism, it implements the prediction of logs that do not appear. To evaluate LogLS, we conducted experiments on two real datasets, and the experimental results demonstrate the effectiveness of our proposed method in log anomaly detection.