K-Prototype Clustering for the Person of Temporary Probation and Parole and Criminal Release Based on Rough Set
Kangshun Li, Ziming Wang, Jiahao Liu, Hongming Fang, Yishu Lei · 2019
Those released from prison, who committed a crime at least once, have higher possibility to commit crimes more seriously. Thus, it's socially significant to reduce the recidivism rate of those people. In this paper, we process the criminals' data obtained from Iowa's public database with two steps. Firstly, Considering the data characteristics of the criminals-many redundant items, we use the rough set attributes reduction algorithm based on probability distribution to reduce the data. Secondly, for clustering the data objects with mixed numeric and categorical attributes, we use the K-prototypes clustering algorithm. Experiment shows the model works well, which could support subsequent prediction work.