Towards Higher Accuracy of Behavioral Big Data Analysis
A Qualitatively Augmented Text Classifier Design Method
A Qualitatively Augmented Text Classifier Design Method
Prof. Dov Te’eni, Coller School of Managemen, Tel Aviv University
Prof. David G. Schwartz, The Social Intelligence Lab, Bar-Ilan University
Dr. Inbal Yahav, Graduate School of Business administration, Bar-Ilan University
This interdisciplinary study develops a hybrid classification method that integrates qualitative analysis with classifier design for text and image analysis of online behavioral big data, which we call Qualitatively Augmented Text Classifier Algorithms (QATCA). This method will be developed and evaluated on an existing cyber intelligence platform that analyzes Dark Web activity undetectably and autonomously. The application of the method relies on computer support that not only aids qualitative analysis and classifier design but also ensures the integration of both components. Efficient and effective classification of online behavioral big data is an important tool for cybersecurity. The output of this study will contribute to the analytical techniques addressing (1) cyber conflict and warfare issues (e.g., monitoring of cyber-crime and cyber-terror activities and the potential criminal’s and terrorist’s behavior); (2) security mechanisms, methodologies, and strategies (e.g., monitoring of online information leakages); and (3) online social networks and subversive behavior (e.g., automatic sentiment trend analysis).