Logo
Nazad
Lamija Zukic, Samed Jukic
0 2021.

Prediction of Solved Homicides Using Classification Method

Homicide rates are still high in the world and they are the worst crime in human existence. Despite all the technological advances and usage of information by various agencies, the number of homicides is not decreasing. Homicide prediction in certain countries should notably be the number one priority, which can help the government to easily identify the kind of profile they are looking for, or even help them prevent those cases. This paper compares different Machine Learning Techniques classifications of homicide prediction. Random Forest (RF), Random Tree, J48, Naive Bayes and k-Nearest-Neighbor (KNN) were tested to determine which method provides the best results in homicide prediction classification. The results of sample accuracy for all algorithms were around 99%, which clearly shows that all algorithms give great results. However, J48 is the best technique applied on the dataset, as it classified all instances correctly.

Pretplatite se na novosti o BH Akademskom Imeniku

Ova stranica koristi kolačiće da bi vam pružila najbolje iskustvo

Saznaj više