Abstract:
Rape is serious crimes that have a troubling impact on victims in particular and the society in
general. It is very dynamic and flexible. It is also sensitive for social and political crises.
Therefore, developing effective rape crime prevention and mitigation methods to protect the
people is indispensable. In relation to these, data mining techniques play a critical role to develop
predictive models that can identify individuals who are at risk of being raped. The trend data
modeling and analysis is essential to interfere and prevent the victim from various suspected
rapes. In this thesis, a predictive model for rape prediction and prevention has developed using
data mining techniques. The K-nearest neighbors (KNN) and decision tree (DT) algorithms have
been used to develop the model. The data obtained from Addis Ababa Police Commission
(AAPC) that has been collected from 11 sub-cities of Addis Ababa City. The data has organized
and prepared in a table form of 10 columns and 4067 rows. The data has been pre-processed and
trained so as to develop the predictive model. After the models have been developed, its
performance has been evaluated using testing data set and accuracy. The KNN and DT models
were able to predict the probability of a rape crime occurring with an accuracy of 93.5% and
88.1% respectively. Based on the result of model comparison score, KNN was more effective
technique than DT. The results show that the system is able to accurately predict the rape crime
in a given area. The models can be used to identify areas where prevention efforts should be
focused and hence to design appropriate interventions mechanisms to prevent rape crime.
Key Words: Rape Crime; Data Mining; K-nearest neighbors (KNN) and Decision Tree (DT).