Crime Analysis Framework for Predicting Criminal Behavioral Patterns with Machine Learning

Data analytics is critical for businesses and organizations to understand their data's behaviour, patterns and insights. Crime analysis is an important component of controlling crime rates in cities and regions based on historical data. This paper proposes a crime analysis methodology supported by machine learning techniques to predict and analyze criminal behaviour patterns. Machine learning algorithms such as Random Forest, Decision Tree, SVM, XGBoost and Multilayer Perceptron are used on the crime dataset for detailed analysis. The framework consists of three phases: data preprocessing, model training, and prediction and analysis. By analyzing crime data using Kaggle's open-source crime records, our project aims to understand crime patterns, determine the crime's severity, and identify where and when it occurred. We use the abovementioned algorithms to classify crime patterns and compare the performance accurately. Our proposed framework uses various visualization techniques to reveal crime trends and predict crimes using machine learning algorithms. Input features used in the algorithms include time, state, district, year, murder, rape, theft and dowry deaths.