Mihai Alexandru Botezatu
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Mihai Alexandru Botezatu has a total of 11 co-authors. Here are the top co-authors by number of shared publications:
- Iustin Priescu
Department of Informatics, Faculty of Informatics, Titu Maiorescu University, 040051 Bucharest, Romania
3 shared publications - Mironela Pirnau
Department of Informatics, Faculty of Informatics, Titu Maiorescu University, 040051 Bucharest, Romania
3 shared publications - Alexandra Hosszu
University of Bucharest
2 shared publications - Alexandru TăbușCă
Romanian-American University
2 shared publications - Catalina Mihaela Priescu
Faculty of Automatic Control and Computer Science, National University of Science and Technology Politehnica Bucharest, 060042 Bucharest, Romania
2 shared publications - Daniela Joita
Department of Informatics, Faculty of Informatics, Titu Maiorescu University, 040051 Bucharest, Romania
2 shared publications
Latest Publications
- Open Access
- 18 April 2026
- Published by MDPI AG in Electronics
- Vol. 15 (8) , 1724
From Vector Space to Symbolic Space: Informational and Semantic Analysis of Benign and DDoS IoT Traffic Using LLMs
This paper investigates the feasibility of using Large Language Models (LLMs) for the structural analysis of flow-based network data. This analysis is carried out in the presence of a structural difference between the multidimensional numerical space of IoT features and the symbolic space in which LLMs operate. The primary objective was the development of a formal framework that enables the controlled transformation of numerical data into linguistically analyzable semantic representations, without resorting to classification or machine learning mechanisms. We propose the Semantic Flow Encoding (SFE) mechanism, a deterministic method for robust discretization and behavioral abstraction that converts the numerical characteristics of Internet of Things (IoT) flows into structural semantic descriptions using the Canadian Institute for Cybersecurity Internet of Things Device Identification and Anomaly Detection (CIC IoT-DIAD) 2024 dataset. Through formal informational measures, it is demonstrated that the existence of an intrinsic structural difference between benign and DDoS traffic in the analyzed dataset. In the validation stage, we evaluated whether these informational differences are reflected at the level of linguistic abstraction through controlled inference experiments in IBM WatsonX. The present paper suggests that LLMs may support semantic auditing of distributional structure when guided by a formal encoding layer. In this manner, a reproducible framework for integrating numerical security data into language-model-based analysis is suggested.
- 18 April 2026
- Published by MDPI AG in Electronics
- Vol. 15 (8) , 1724
Open AccessFrom Vector Space to Symbolic Space: Informational and Semantic Analysis of Benign and DDoS IoT Traffic Using LLMs
This paper investigates the feasibility of using Large Language Models (LLMs) for the structural analysis of flow-based network data. This analysis is carried out in the presence of a structural difference between the multidimensional numerical space of IoT features and the symbolic space in which LLMs operate. The primary objective was the development of a formal framework that enables the controlled transformation of numerical data into linguistically analyzable semantic representations, without resorting to classification or machine learning mechanisms. We propose the Semantic Flow Encoding (SFE) mechanism, a deterministic method for robust discretization and behavioral abstraction that converts the numerical characteristics of Internet of Things (IoT) flows into structural semantic descriptions using the Canadian Institute for Cybersecurity Internet of Things Device Identification and Anomaly Detection (CIC IoT-DIAD) 2024 dataset. Through formal informational measures, it is demonstrated that the existence of an intrinsic structural difference between benign and DDoS traffic in the analyzed dataset. In the validation stage, we evaluated whether these informational differences are reflected at the level of linguistic abstraction through controlled inference experiments in IBM WatsonX. The present paper suggests that LLMs may support semantic auditing of distributional structure when guided by a formal encoding layer. In this manner, a reproducible framework for integrating numerical security data into language-model-based analysis is suggested.
- Open Access
- 16 March 2026
- Published by MDPI AG in Preprints.org
From Vector Space to Symbolic Space: Informational and Semantic Analysis of Benign and DDoS IoT Traffic Using LLMs
In this paper, we investigate the feasibility of using Large Language Models (LLMs) for the structural analysis of flow-based network data, considering the fundamental onto-logical difference between the multidimensional numerical space of IoT data and the symbolic space in which these models operate. The primary objective was the development of a formal framework that enables the controlled transformation of numerical data into linguistically analyzable semantic representations, without resorting to classification or machine-learning mechanisms. We propose the SFE mechanism, a deterministic method for robust discretization and behavioral abstraction that converts the numerical characteristics of IoT flows into structural semantic descriptions, based on the CIC IoT-DIAD 2024 [1] dataset. Through formal informational measures, we demonstrate the existence of an intrinsic structural difference between benign and DDoS traffic in the analyzed dataset. In the validation stage, we evaluated whether these informational differences are reflected at the level of linguistic abstraction through controlled inference experiments in IBM WatsonX [2]. The paper demonstrates that LLMs can work as mechanisms for semantic auditing of distributional structure when supported by a formal encoding layer, offering a reproducible framework for integrating numerical security data into language-model-based analysis.
- 16 March 2026
- Published by MDPI AG in Preprints.org
Open AccessFrom Vector Space to Symbolic Space: Informational and Semantic Analysis of Benign and DDoS IoT Traffic Using LLMs
In this paper, we investigate the feasibility of using Large Language Models (LLMs) for the structural analysis of flow-based network data, considering the fundamental onto-logical difference between the multidimensional numerical space of IoT data and the symbolic space in which these models operate. The primary objective was the development of a formal framework that enables the controlled transformation of numerical data into linguistically analyzable semantic representations, without resorting to classification or machine-learning mechanisms. We propose the SFE mechanism, a deterministic method for robust discretization and behavioral abstraction that converts the numerical characteristics of IoT flows into structural semantic descriptions, based on the CIC IoT-DIAD 2024 [1] dataset. Through formal informational measures, we demonstrate the existence of an intrinsic structural difference between benign and DDoS traffic in the analyzed dataset. In the validation stage, we evaluated whether these informational differences are reflected at the level of linguistic abstraction through controlled inference experiments in IBM WatsonX [2]. The paper demonstrates that LLMs can work as mechanisms for semantic auditing of distributional structure when supported by a formal encoding layer, offering a reproducible framework for integrating numerical security data into language-model-based analysis.
- 4 April 2025
- Published by IGI Global Scientific Publishing
- p. 217-248
Sustainable Development and the Attractiveness of Development Regions
The study demonstrates, through empirical research on the development regions in Romania, that the average monthly income per household can be an element for evaluating the attractiveness of the regions for an adult, especially a young person, who is looking for a suitable job or an area in which to settle down and build a future. We considered factors that might be correlated with economic development in the region (GDP/place at the regional level, population occupancy rate, investments, turnover of active companies, tourist accommodation capacity, the employed population with higher education in the region). We built simple and multiple linear regression models, with data series from 2007-2019, obtained by processing data from the Statistical Yearbook of Romania (2008-2020). We concluded that all these factors are correlated with the size of the average monthly income indicator on a regional household. The results were interpreted and explained.
- 4 April 2025
- Published by IGI Global Scientific Publishing
- p. 217-248
Sustainable Development and the Attractiveness of Development Regions
The study demonstrates, through empirical research on the development regions in Romania, that the average monthly income per household can be an element for evaluating the attractiveness of the regions for an adult, especially a young person, who is looking for a suitable job or an area in which to settle down and build a future. We considered factors that might be correlated with economic development in the region (GDP/place at the regional level, population occupancy rate, investments, turnover of active companies, tourist accommodation capacity, the employed population with higher education in the region). We built simple and multiple linear regression models, with data series from 2007-2019, obtained by processing data from the Statistical Yearbook of Romania (2008-2020). We concluded that all these factors are correlated with the size of the average monthly income indicator on a regional household. The results were interpreted and explained.
- Open Access
- 31 January 2024
- Published by MDPI AG in Electronics
- Vol. 13 (3) , 584
Content Analysis Using Specific Natural Language Processing Methods for Big Data
Researchers from different fields have studied the effects of the COVID-19 pandemic and published their results in peer-reviewed journals indexed in international databases such as Web of Science (WoS), Scopus, PubMed. Focusing on efficient methods for navigating the extensive literature on COVID-19 pandemic research, our study conducts a content analysis of the top 1000 cited papers in WoS that delve into the subject by using elements of natural language processing (NLP). Knowing that in WoS, a scientific paper is described by the group Paper = {Abstract, Keyword, Title}; we obtained via NLP methods the word dictionaries with their frequencies of use and the word cloud for the 100 most used words, and we investigated if there is a degree of similarity between the titles of the papers and their abstracts, respectively. Using the Python packages NLTK, TextBlob, VADER, we computed sentiment scores for paper titles and abstracts, analyzed the results, and then, using Azure Machine Learning-Sentiment analysis, extended the range of comparison of sentiment scores. Our proposed analysis method can be applied to any research topic or theme from papers, articles, or projects in various fields of specialization to create a minimal dictionary of terms based on frequency of use, with visual representation by word cloud. Complementing the content analysis in our research with sentiment and similarity analysis highlights the different or similar treatment of the topics addressed in the research, as well as the opinions and feelings conveyed by the authors in relation to the researched issue.
- 31 January 2024
- Published by MDPI AG in Electronics
- Vol. 13 (3) , 584
Open AccessContent Analysis Using Specific Natural Language Processing Methods for Big Data
Researchers from different fields have studied the effects of the COVID-19 pandemic and published their results in peer-reviewed journals indexed in international databases such as Web of Science (WoS), Scopus, PubMed. Focusing on efficient methods for navigating the extensive literature on COVID-19 pandemic research, our study conducts a content analysis of the top 1000 cited papers in WoS that delve into the subject by using elements of natural language processing (NLP). Knowing that in WoS, a scientific paper is described by the group Paper = {Abstract, Keyword, Title}; we obtained via NLP methods the word dictionaries with their frequencies of use and the word cloud for the 100 most used words, and we investigated if there is a degree of similarity between the titles of the papers and their abstracts, respectively. Using the Python packages NLTK, TextBlob, VADER, we computed sentiment scores for paper titles and abstracts, analyzed the results, and then, using Azure Machine Learning-Sentiment analysis, extended the range of comparison of sentiment scores. Our proposed analysis method can be applied to any research topic or theme from papers, articles, or projects in various fields of specialization to create a minimal dictionary of terms based on frequency of use, with visual representation by word cloud. Complementing the content analysis in our research with sentiment and similarity analysis highlights the different or similar treatment of the topics addressed in the research, as well as the opinions and feelings conveyed by the authors in relation to the researched issue.
- 22 September 2023
- Published by Bucharest University of Economic Studies in ECONOMIC COMPUTATION AND ECONOMIC CYBERNETICS STUDIES AND RESEARCH
- Vol. 57 (3/2023) , 133-150
Economic Factors Correlated with Marriage and Divorce Rates in European Union Countries
- 22 September 2023
- Published by Bucharest University of Economic Studies in ECONOMIC COMPUTATION AND ECONOMIC CYBERNETICS STUDIES AND RESEARCH
- Vol. 57 (3/2023) , 133-150
Economic Factors Correlated with Marriage and Divorce Rates in European Union Countries