Aisultan Shoiynbek

The rapid migration of financial fraud to encrypted messenger platforms such as WhatsApp and Telegram has created an urgent need for lightweight, endpoint-based detection tools capable of operating on small, multilingual datasets characteristic of real-world "cold start" conditions. Unlike traditional email spam, messenger-based fraud is short-lived, conversational, and heavily obfuscated through mixed scripts and artificial urgency, rendering static rule-based filters ineffective. Existing machine learning solutions predominantly address English-language SMS spam and rely on large labeled corpora, leaving a significant gap for post-Soviet and Central Asian deployment contexts where Russian, Kazakh, and English frequently co-occur within a single message. This study aims to identify the optimal classical machine learning architecture for binary fraud classification of such multilingual text messages under severe data scarcity constraints. A balanced pilot dataset of 189 messages (95 legitimate / 94 fraudulent) was manually collected from Telegram, WhatsApp, Reddit, Pikabu, and verified news sources. Text was vectorized using TF-IDF (unigrams and bigrams, max 100 features, min_df=2) without morphological preprocessing to preserve cross-language token integrity. Five classifiers were benchmarked: Logistic Regression, Linear SVM, Multinomial Naive Bayes, Random Forest, and XGBoost. Regularization strategies including L2 penalty, additive smoothing, and structural tree constraints, combined with class-weight balancing, were applied to mitigate overfitting on the small sample. Model evaluation used 5-fold stratified cross-validation and an independent test set, with Recall (Fraud) as the primary safety metric alongside Accuracy, F1-score, and ROC-AUC. Logistic Regression and Linear SVM achieved the highest overall performance (Accuracy = 0.97, Recall = 1.00, F1 = 0.97, ROC-AUC = 0.957–0.962), successfully identifying all fraudulent messages with only one false alarm each. XGBoost also reached perfect recall (1.00) at a cost of lower precision (Precision = 0.88), while Random Forest and Naive Bayes were less stable, achieving accuracies of 0.86 and 0.79 respectively. The findings demonstrate that for small, sparse, and multilingual fraud datasets, aggressively regularized linear models with minimal feature spaces outperform complex ensemble architectures, offering a viable, computationally lightweight foundation for real-time endpoint fraud filtering on encrypted platforms.

The expansion of telecommunication services has been met with a rise in the cases of the fraudulent phone calls posing a big threat to individuals and organizations. Conventional techniques of detecting are usually based on offline analysis of full conversations, which restricts their promptness of intervention. In this paper, the author proposes a real-time, turn-taking, fraud detecting system, which is based on pre-trained contextual embeddings in combination with a bi-directional Long Short-Term Memory network in order to model semantic content and temporal dynamics of multi-turn conversations. To detect fraudulent calls, the system progressively changes the probability of a call being a fraud after every conversational turn to allow it to detect a fraud. When tested with a synthetic multi-turn dialogue dataset, it is shown that the proposed BiLSTM using BERT embeddings has a test accuracy of 93.75% and an F 1 score of 93.55, which is higher than the current machine learning and convolutional baselines. The system can note most of the scams during the initial few turns of a call, which offers fast risk evaluation. These findings suggest the usefulness of context-based, progressing modeling to detect fraud in real time and its possibility of practical application.

The increasing number of messenger fraud cases requires early and precise threat detection at unprecedented levels. The research examines modern NLP-based approaches which detect deceptive messages in messaging applications. The research examines various NLP approaches which analyze text data from different messaging platforms through text classification and tonality analysis and anomaly detection and thematic modeling techniques. The paper examines model learning data types together with text pre-processing methods and essential text features and evaluates traditional methods (e.g., Bag of Words, TF-IDF) and modern neural networks. The researchers encounter multiple obstacles while working which include the complex nature of processing informal language and the presence of noisy data and the need to frequently update models to detect new fraudulent schemes. The research focuses on messenger platform fraud detection because it addresses the unique challenges of real-time message streams and informal language and multimodal communication. The review evaluates technical and contextual aspects by presenting suitable models and architectures for dynamic short-form content and identifying technologies that deliver low-latency responses. The research aims to assess existing methods while identifying optimal approaches and proposing new directions to boost the accuracy and reliability of messenger fraud detection systems.

With the increasing prevalence of mental health issues, particularly suicidal behaviors, the need for early and accurate detection has become critical. This paper explores the current landscape of machine learning approaches used for the detection of suicidal tendencies. It examines a wide range of machine learning techniques applied to various data sources, including social media, clinical records, psychological assessments, self-reported forms like PHQ-9, audio speech recordings, and multimodal data integrating speech and visual information. This comprehensive review aims to reveal the types of existing research based on these varied datasets, highlighting the nuances of data collection, significant features identified, and the results obtained by different studies. Additionally, the review discusses the challenges and limitations associated with these approaches, providing researchers and practitioners with valuable insights into the potential and pitfalls of machine learning applications in diagnosing individuals at risk of suicide. The goal is to inform future research and improve early detection methods to ultimately reduce suicide rates.

This paper evaluates and compares the performances of three well-known optimization algorithms (Adagrad, Adam, Momentum) for faster training the neural network of CTC algorithm for speech recognition. For CTC algorithms recurrent neural network has been used, specifically Long-Short-Term memory. LSTM is effective and often used model. Data has been downloaded from VCTK corpus of Edinburgh University. The results of optimization algorithms have been evaluated by the Label error rate and CTC loss.