A Proposed Big Data Analytics Model for Crimes Predication based on Spatial and Temporal Criminal Hotspot

Big data is simply large, voluminous data collected from different sources which could be structured or unstructured. The ancient processing system may not be successful in processing such voluminous data. Big data analytics (BDA) uses extensive techniques and tools for analyzing large. This Paper introduced a solution to the crime prediction problem using machine learning algorithms. to classify this monstrous data. SVM and K-NN, Random Forest, and MLP algorithms solve this problem. In this Paper addresses, the researcher finds the best SVM algorithm, the main thing to keep in mind here is that these are to give you the most accurate answer possible as quickly as possible. The Predictive model shows that the prophet model handles outliers well and it is robust to missing data and shifts in the trend. These outcomes are going to benefit the police departments to better understand crime issues and provide insights that will enable them to track activities, predict the likelihood of incidents, effectively deploy resources, and optimize the decision-making process. After applying the proposed model to many machine learning algorithms, the paper concluded that the best crime prediction algorithm in terms of accuracy is in order as follows: SVM algorithm with an accuracy of 0.99997, time 1100 seconds, followed by K-NN algorithm with an accuracy of 0.999976, time 2450 seconds, then Random Forest algorithm with accuracy 0.999996, Time 1420 seconds, then MLP algorithm with accuracy 0.995886, time 1346 seconds. The best algorithm to achieve the best accuracy in the least time is SVM.