Title: Predicting sexual offenders using exhaustive CHAID techniques on victim's age

Authors: Bhajneet Kaur; Laxmi Ahuja; Vinay Kumar

Addresses: Amity University, Noida (RDIAS), Uttar Pradesh, India ' Amity Institute of Information Technology (AIIT), Amity University, Noida, Uttar Pradesh, India ' Computer Society of India (CSI) Delhi Chapter, Delhi, India

Abstract: Sexual offences can spoil the whole culture of the society. This research paper proposes two decision models to classify and predict the sexual offenders of minor and major victims on the basis of their physical attributes namely age, race, weight and height using CHAID and Exhaustive CHAID techniques of decision tree. Overall dataset has been divided into 70:30 for building and testing the models. As resulted 79.8% rate of accuracy found by model using CHAID technique even model tested with 79.1% rate of accuracy. By using Exhaustive CHAID, 79.9% rate of accuracy depicts by the model developed through 70% of test data and model validated through 30% of test data with 78.8% rate of accuracy. The proposed models can help to take any kind of decision further by police departments, sexual harassment cells and law enforcement agencies for security purposes.

Keywords: sexual offenders; minor victim; major victim; decision tree; index values; gain chart; response chart; machine learning; CHAID; SPSS; exhaustive CHAID.

DOI: 10.1504/IJAMS.2022.121043

International Journal of Applied Management Science, 2022 Vol.14 No.1, pp.71 - 89

Received: 03 Apr 2019
Accepted: 28 Sep 2019
Published online: 23 Feb 2022 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article