The process of feature extraction of sports data has strong fuzziness and weak correlation.Differ-ent from other data,the multiple indicators of sports data show a linear correlation process,but this feature is not used by the current classical fusion algorithm,making the fusion coverage low.A method of sports data feature fusion based on linear fitting is designed.Firstly,a feature extraction model is built,and then the updated rules of data feature extraction are adopted through the monitoring system.Nextly,the extracted data features are counted and denoised.Finally,regression analysis and linear fitting methods are used to quantify the features thus completing the sports data feature fusion.The results show that the proposed method has high coverage and high reliability.Under different levels of Gaussian noise,the SNR is high,the extraction effect is good,the energy consumption is low,and the fitting degree is high.
sportsfuzzinessphysiological index datafeature extractionlinear fitting
NETL
NSTL
万方数据
