A Comparison of Sensitive Information Detection Framework using LSTM and RNN Techniques

Abstract

Sensitive information is meant to be stored securely to avoid any data breach. Hence this research emphasized on whether or not Long Short Term Memory and Recurrent Neural Network is suitable for sensitive information detection framework since performance analysis on this method are not common. In this research, the objectives are to design Sensitive Information Detection Framework using Deep Learning techniques, to detect sensitive information using LSTM and RNN techniques and to test and validate the sensitive information detection framework performance. This research uses raw data set given from Book Hack Enterprise company and by using Weka Explorer software tool to go through 5 phases in the proposed framework, which are Dataset, Preprocessing, Feature Extraction, Classification Algorithm and Performance Evaluation. This research evaluates the model performance based on Accuracy, Precision, Recall and F1-score.