MODEL OF A SOFTWARE ENVIRONMENT DEFECT DETECTION SYSTEM BASED ON DEEP LEARNING WITH THE MOST SUITABLE HYPERPARAMETERS

Potapov Dmitry Artemovich, Kornienko Svetlana Vladimirovna

This article discusses methods for detecting errors in software code, lists existing tools for searching for defects in the software environment, focuses on a way to find and predict the appearance of software code defects using machine learning, lists well-known scientific works on finding errors and anomalies in the software environment, considers a typical scheme of artificial neural network operation in training A hybrid model of a software environment defect detection system based on deep learning with the most appropriate hyperparameters is proposed, which is based on machine learning technology with hyperparameters calculated using the hierarchy analysis method.

Keywords:machine learning, software environment defects, hyperparameters, hybrid model, hierarchy analysis method