In predicting the new crime categories, average property value was the most important variable in the bagging and random forest models, and second most important in the neural network model. We also observe that college and university distance is most important in the neural network model and second most important in the random forest model. We observed earlier that for the original categories, though we were predicting different outcome variables, average property value was still one of the most important predictors. This could indicate a discrepancy between prevalent crime types based on property values, which is in turn an indicator of wealth and income disparity. Furthermore, discrepancies in crime type based on proximity to colleges and universities could signal to law enforcement what to prioritize near college and university facilities.

Home

Sources

Jupyter Notebook