Information and communication and chemical technologies

Vol. 4 No. 29 (2025) / 2025-12-30 / Number of views: 192

DEVELOPMENT OF A DEEP NEURAL NETWORK MODEL FOR FORENSIC ANALYSIS ON TEXTUAL DATA

Authors

L.N. Gumilyov Eurasian National University
https://orcid.org/0009-0002-3377-7003
L.N. Gumilyov Eurasian National University
https://orcid.org/0000-0001-5152-147X
Astana IT University
https://orcid.org/0000-0002-5715-4954
Russian Academy of Sciences, Federal Research Center
https://orcid.org/0000-0002-7935-0569

Keywords

text data, forensics, artificial intelligence, deep learning, LSTM, Attention, linguistic analysis

Link to DOI:

https://doi.org/10.58805/kazutb.v.4.29-1243

How to quote

Kuttybek, A.A. Kuttybek, et al. “DEVELOPMENT OF A DEEP NEURAL NETWORK MODEL FOR FORENSIC ANALYSIS ON TEXTUAL DATA”. Vestnik KazUTB, vol. 4, no. 29, Dec. 2025, doi:10.58805/kazutb.v.4.29-1243.

Abstract

This paper explores the use of artificial intelligence and deep neural network technologies for the intelligent analysis and automated forensic examination of textual data. Today, the rapid growth of information on the Internet and the widespread use of social networks pose significant challenges for data processing and analysis by law enforcement agencies. During investigations and forensic examinations, digital evidence increasingly contains textual materials; therefore, their automatic processing and forensic evaluation have become a relevant area of research.

The main objective of this study is to develop a deep neural network model capable of detecting signs of offenses, as well as aggressive or extremist elements in textual data. The proposed model is based on the LSTM architecture combined with the Attention mechanism, which allows efficient analysis of the semantic structure, contextual relationships, and emotional tone of texts in the Kazakh language.

At the preprocessing stage, the data undergo tokenization, noise and symbol removal, and vectorization using the TF-IDF and Word2Vec methods. Based on the prepared dataset, the model is trained and achieves a classification accuracy of up to 92% for forensic-related texts. This result outperforms classical algorithms such as Logistic Regression and SVM.

The proposed approach enhances both the speed and accuracy of evidence analysis in forensic examinations. It represents an important step toward automating investigations, facilitating the work of forensic experts, and reducing human error. The results obtained can be applied in the further development of intelligent expert systems, the expansion of the Kazakh-language forensic text corpus, and the improvement of artificial intelligence applications in the field of law enforcement.