AUTOMATED SYSTEM FOR GENERATIVE MULTIMODEL MULTICLASS CLASSIFICATIONS AND SIMILARITY ANALYSIS USING MACHINE LEARNING

摘要

To provide an automated system for generative multimodel multiclass classifications and similarity analysis which efficiently makes classifications and similarity analysis using probabilistic models generated through machine learning thereby capable of indicating whether a file is likely to include a malicious code.SOLUTION: A system extracts features from a sample of data, makes classifications regarding some or all of the extracted features, and generates machine learning models using such data that can be used to characterize data samples subsequently. The system makes classifications and similarity analysis of samples using models thereby determining if a specific computer file is likely to be malicious.SELECTED DRAWING: None