Predictive Data Mining Model for Cyber Crime

Abstract

Application of data mining techniques (DMT) in different field like e-governance, e-agriculture, e-health, e-education, etc has contributed significant results. The results produced by DMT can be utilized by policy maker/ government officer for people welfare and making good governance. One of the applications of DM may be cyber crime classification. Internet usage has increased tremendously since the last few years, due to which the incidents of cybercrime have also increased rapidly. It causes the generation of new dataset in the field of cyber crimes. For these dataset, efficiency of the data mining algorithm is yet to be checked. Hence, this paper deals with the objective to develop efficient classification model to classify cyber crime incidents. We propose an iterative classification model using data mining techniques. Five classification models Naïve Bayes, Support Vector Machine, K Nearest Neighbors, Decision List and Decision Tree have been trained and tested on selected dataset. The iterative classification model relation with SVM was found best with 95.03% accuracy as compare to other classification models.