Enhancing automatic number plate recognition accuracy with novel support vector machine algorithm and comparison with Lasso Regression

In this study, we used the Support Vector Machine (SVM) technique to build a system that could recognise and identify licence plates, and we compared its accuracy to that of the Lasso Regression approach to see how well it performed (TM). The Parts and Methods: Using machine learning techniques like the Novel Support Vector Machine Algorithm and the Lasso Regression algorithm, the methodology employs a sample size of five hundred and fifty datasets sourced from IEEE-dataport.org, with twenty datasets used for each group. The sample size was determined with the following parameters: G power = 80%, alpha = 0.05, and confidence level = 96%. Findings: The Novel Support Vector Machine Algorithm outperformed the Lasso Regression algorithm with an accuracy of 88.99 percent ("Analysis and Comparison for Innovative Prediction Technique Using Logistic Regression Algorithm over Support Vector Machine Algorithm with Improved Accuracy" 2022). Both algorithms were compared to one another. Utilizing independent sample t-tests, it was shown that the two methods varied significantly in terms of accuracy. The groups were found to be statistically significant, since the p-value for these tests was 0.0057, which is greater than 0.05. The Novel Support Vector Machine Technique outperformed the Lasso Regression technique by a wide margin, with an accuracy of 88.99 percent compared to 84.82 percent.