Modelo de clasificación para la identificación de software malicioso ofuscado en sistemas operativos windows

Obfuscation is a technique in computer science that makes it difficult to understand source code in order to protect intellectual property and prevent reverse engineering; in the scope of this study, it is used to complicate malware detection. However, to address this problem, a model based on machine learning techniques was built to identify threats that use obfuscation techniques. Therefore, in the development of this study, the “Knowledge Discovery in Databases” (KDD) methodology was used, which began with the preparation of the data set, where attribute selection based on correlation was applied. Through a literature review, supervised techniques were selected and applied in the data mining phase. The Random Forest, Decision Tree, SVM, KNN and Gradient Boosting algorithms were used to correctly identify the main groups of malicious software, thus demonstrating the performance of the model in identifying malware. Finally, the main contribution of this research is a model based on the Random Forest algorithm that presented a 99% accuracy in the classification of obfuscated malware, improving the capabilities of identifying cyber threats in this area

• 1. Verfasser: Ponce Velez , Iván Darío (author)
• Format: masterThesis
• Sprache: spa
• Veröffentlicht: 2024
• Schlagworte: Malware ofuscado; Aprendizaje automático; Ciberseguridad; Técnicas supervisadas
• Online Zugang: https://repositorio.uteq.edu.ec/handle/43000/7844