Revealing Insights into Criminal Behaviour: Exploring Patterns and Trends Through Machine Learning Predictive Models

This research paper explores the feasibility of machine learning models in predicting human criminal behavior based on historical and collected datasets. The study design involves data collection from public records, past criminal records, and evidence such as audio and video materials. The collected data is segregated according to specific requirements for analysis. Analytical techniques employed in this research include linear regression, K-Nearest Neighbors (KNN), Support Vector Machines (SVM), RNN, random forests, logistic regression, and Autoencoders. Through extensive experimentation on diverse datasets amounting to 15.7GB, the proposed models yielded an overall prediction accuracy of 83.6%. The study contributes to the development of an artificial criminal behavior deduction model. However, it is essential to acknowledge the complexity and ethical implications of predicting human behavior. The paper emphasizes the need for caution in interpreting results and highlights potential biases and uncertainties. This research strives to present a comprehensive analysis of machine learning's potential in understanding human criminal mindsets, raising awareness about the capabilities and limitations of current predictive models. It underscores the significance of responsible and ethical use of machine learning for sensitive applications in the criminal justice domain.