Analyzing Factors Influencing Crime Rates in Communities by Lasso Regression

Effective crime rate prediction is essential for public safety and urban planning. This study develops a predictive model using Lasso regression to identify key factors influencing crime rates while addressing noise variables. Our model highlights five significant predictors: median housing value, median owner cost as a percentage of income, median rent, percentage of people with a bachelor’s degree or higher, and percentage of teenagers in two-parent households. Higher housing values and rents correlate with lower crime rates, while higher housing costs relative to income correlate with increased crime. Education levels and stable family structures also contribute to lower crime rates. This paper outlines the data cleaning, model evaluation, and the importance of addressing noise variables for robust predictive modeling.