MVLSC Home · Issue Contents · Forthcoming Papers
The Application of Improved HMM Training Detection Algorithm in the Improvement of Dragon Boat Sports Skills
Meizhi Wang
With the rapid development of artificial intelligence and machine learning technology, its application scope is expanding day by day, but its application in dragon boat movement is very limited at present. Based on this, the GA-BW algorithm is obtained by improving the HMM training detection algorithm, and the relevant model is constructed to achieve the effective recognition of HMM model in dragon boat training detection, and its effectiveness and accuracy are verified. The experimental results show that the overall recognition rate of GA-BW model is high, and the recognition rate can reach 93.8%, and the highest recognition rate can be achieved when the number of HMM model states is 5. In addition, in the threshold limit experiment, it is found that the increasing threshold also reduces the error rate of the algorithm and improves the detection rate. To sum up, in the training and detection of HMM model, GA-BW algorithm can be fully used to accurately describe the relevant movements of athletes, and through the analysis of sports data, it can effectively help athletes correct action errors and improve their technical level. In the actual dragon boat training, the athletes can carry out targeted training to complement their own weaknesses and improve the skills of dragon boat sports, which is highly practical.
Keywords: Training detection algorithm, dragon boat sport, sports skills, probability distribution, recognition rate