Date of Award

8-2024

Document Type

Project

Degree Name

Master of Science in Computer Science

Department

School of Computer Science and Engineering

First Reader/Committee Chair

Dr.Yan Zhang

Abstract

This project employs machine learning methods like K Nearest Neighbors (KNN), Random Forest, Logistic Regression, and Decision Tree algorithms to monitor crime data based on location and pinpoint areas with risks. The project implements and tunes the four models to improve the precision of predicting crime levels. These models collaborate to offer a trustworthy evaluation of crime patterns. K Nearest Neighbors (KNN) categorizes locations by examining the proximity of data points considering coordinates and other factors to identify trends linked to increased crime data. Logistic Regression gauges the likelihood of crime incidents by studying the connection, between factors (like location and time ) and the crime activity, assisting in forecasting crimes in various regions. Decision Tree Classifier uses a tree structure to make decisions based on feature values dividing the data into branches representing decision paths. This approach is particularly useful for identifying high-risk areas using crime data. Random Forest Classifier constructs decision trees and combines their results for classification purposes, resulting in enhanced prediction accuracy and robustness by merging outcomes from multiple trees, thus reducing the risks of overfitting and improving generalization to unseen data.

The system’s efficiency is assessed using a crime dataset that includes information, about crime occurrences, geographical locations, and time-related data. Metrics, like accuracy, precision, and recall are employed to assess the model’s ability to anticipate crimes and identify hotspots accurately.

Share

COinS