Skip to content

Latest commit

 

History

History
75 lines (46 loc) · 2.86 KB

README.md

File metadata and controls

75 lines (46 loc) · 2.86 KB
Raw

Juvenile Crime Analysis

Using data exploration, data analysis and machine learning.

Objective

Main objective of the analysis is prediction and interpretation of juvenile crime in India and deciding efficient allocation of resources in the right areas to minimise the crime rate.

Dataset

Juvenile Crime in India

Dataset we have considered:Github

Crime in India (Kaggle)

Following link contains processed data of various crime in separate csv file. From which we have considered csv which contains information about juvenile crime:Kaggle

Crime in India (NCRB)

Following link contains the original unprocessed data of various crime in separate csv file:NCRB

Metadata

The dataset contains different sub datasets, each of these sub datasets holds the aggregate value of crime recorded from 2001 to 2010 and are split on various categories. On sub-dataset divided based on:

  • Age-group and type of crime
  • Background
  • Recidivism The kaggle dataset was preprocessed, various combinations of datasets was used to train the model and address various issues.

Problem Statements And Results

Given state, crime(mapping required, not considered for now), age, gender, predicting the possibility of SLL(Naive bias using above data + ARIMA to predict current year using previous year data)

SLL by parima

Alt text

Naive Bayes Result

Boys

Alt text

Girls

Alt text

Feature Importance(Using Linear Regression and Random Forest)

Relation between punishment and recidivism (Linear Regression)

Alt text Alt text

Relation between punishment and recidivism (Random Forest)

Alt text Alt text

Background of juveniles committing crime

Alt text Alt text

Which State requires better rehabilitation facilities.(Done using clustering)

Alt text Alt text

Given State, who to be on a look out for?(Arima+Naive bias, 3 class classification)

Recidivism by parima
Alt text

Naive Bayes Result
Alt text

Future Scope

The next step is to streamline the data analysis process by improving the collection method of data or aggregation of various data. Better interpretations can be made with more abstract data. A lot of key information is lost in the current format of data aggregation.