Cricket's Computational Core: Enhanced Shot Classification with Convolutional Neural Networks And SVM

In the abstract of the research paper titled “Enhancing Cricket Shot Identification Using the CNN-SVM Approach: In our paper “Five-Class Analysis,” we introduce a novel model that combines the positive features of convolutional neural networks (CNN) and support vector machines (SVM) to detect five different classes of cricket shots: cover drive, cut shot, pull shot, leg glance, and straight drive. The model has been carefully evaluated using a solid set of metrics, allowing the results to reflect good accuracy across all classes: 87.85% in leg shots and 90.433% in pull shots,924 L cut shots, and deep straight drives, respectively. Considering leg glance, this value goes as high as. The results are equally impressive, with scores of 96.36%, 95.15%, and 87.43%, and including other sub-scores amounting to an overall accuracy that comes close to the analysis that is enhanced by using macro, micro, and weighted averages to provide an integral evaluation of the model's predictive capability. The macro-averaged scores show that the classification performance is balanced, with precision and recall of 91.77% and F... The weighted averages corrected for class imbalance exhibit an accuracy of 92.64%, a recall of 93.58%, and a score of 92.56%. The micro, which records the total contribution of all classes' accuracy rate, is 92.58% across precision curve recall and F score, displaying consistency to a high level in performance. These statistical values indicate the effectiveness of the CNN-SVM model in classifying complex cricket shots and emphasize its potential as a valuable tool for analytics sports technology.