ABSTRACT
Prisoner coaching will be more effective if factors relating to the causes of crime are classified. Statistical analysis is used to classify the prisoners’ socioeconomic factors. The multinomial logistic regression method is suitable for this study because the response variable is in more than two categories. In this study, the response variables were narcotics crime, fraud, theft, and other crimes. The explanatory variables that affect the response variable are socioeconomic. The data relates to inmates in DKI Jakarta in 2018 using 600 respondents. Based on the odds ratio with a significant level of 0.05, it is known that 11 explanatory variables together have a substantial influence on the type of crime associated with inmates in DKI Jakarta. Variables that partially affect the crime of theft are age, income, expenditure, communication tools, and homeownership. Influencing variables for narcotics crime are religion, age, education, income, expenditure, dependency burden, course, and communication tools, while some of the variables that influence other crimes are religion, age, education, income, expenditure, dependency, and homeownership.
