A Comparative Study of Decision Review System in Cricket: Residual Network Vs SVM Algorithm Poor Accuracy in Decision Review System in Cricket

This study evaluates and compares the performance of Support Vector Machine (SVM) and Residual Network (ResNet) algorithms within the Cricket Decision Review System (DRS) to enhance umpiring accuracy. Using a dataset of 3400 analyses, we implemented both algorithms for predicting online video quality and assessed their predictive accuracy with a sample size of 10. The results demonstrated that the Standard SVM achieved an impressive accuracy of $\mathbf{9 9 \%}$, significantly outperforming the Residual Network, which achieved only $59 \%$. Statistical analysis confirmed that this difference in accuracy was significant, with less than $5 \%$ probability of occurrence by chance. The findings indicate that SVM offers a notable improvement in average accuracy compared to ResNet, highlighting its potential to refine umpiring decisions and positively impact the sport of cricket.