The possibility of detecting false positive cybersecurity incidents using deep learning models - GRU, Bidirectional LSTM (Bi-LSTM), LSTM - has been studied. The results obtained demonstrate the effectiveness of solving the problem for Powershell scripts. The Bi-LSTM model showed the best classification results, demonstrating an accuracy of 98.50% on the test sample.
Keywords: machine learning, classification, cybersecurity, deep learning, Powershell