Crime will continue to be a big concern to all groups and peoples around the globe, and the complexity of technology and processes that are being exploited to allow for incredibly intricate criminal activities will continue to be a significant threat as well. In order to identify and analyse crime patterns and trends, a rigorous process known as criminal analysis is used. As the usage of digital technologies grows, crime data analysts may be able to aid law enforcement officers in speeding up the investigation process. It's because of this that data mining has become a significant weapon in the fight against crime. Governments and corporations throughout the globe are using it to unearth hidden information from enormous volumes of data, and it is regulated by both. Unstructured information may be mined to uncover previously unknown, useful information, making the system more efficient. Police departments utilise analytical and predictive methods to detect offenders in predictive policing. It has been shown to be really helpful in this area.. To keep up with the ever-increasing volume of crime data, the system will have to automate the evaluation of vast amounts of data that are now stored in warehouses.Additionally, criminals are becoming more good with technology, thus it is vital to deploy new technology in order to keep the police one step ahead of them. Comparisons between deep learning and machine learning for crime prediction are made in this research. This article explains how to extract and classify features for crime pattern analysis. It is possible to begin by preprocessing the input data, and then utilising Deep component level set analysis to identify the most important properties. Gaussian naive bayes classifier and deep learning based multilayer perceptron may be utilised to classify the criminal patterns in this instance. Kaggle's standard crime data has been used in the tests.. Deep learning-based prediction model outperforms machine learning-based model, according to the results.

Crime, Deep component level set analysis, gausian naïve bayes, deep neural network multilayer perceptron, kaggle database