Abstract:
In order to improve the accuracy of temperature prediction for poultry houses and reduce the impact of data redundancy and difference on the prediction results, a temperature prediction model based on intelligent optimization feature subset selection and fuzzy clustering improved SVR(Support Vector Regression) is proposed. Firstly, the optimal feature subset selection model is constructed, and the optimal feature subset selection index is designed to reduce the redundancy and data dimension between features; The improved discrete Gray Wolf algorithm is used to solve the feature subset selection model to realize the optimal feature subset selection. Secondly, the fuzzy clustering improved SVR prediction mechanism is established, and the multi-core FCM(Fuzzy C-means) algorithm is designed to realize the automatic classification of data samples; A SVR prediction algorithm corresponding to data sample classification is proposed, and the Gray Wolf algorithm is used to optimize the SVR parameters to minimize the impact of sample data differences on prediction accuracy. Finally, the optimal feature subset selection and fuzzy clustering are combined to improve the SVR prediction mechanism to realize the high-precision prediction of poultry house temperature. The simulation results show that the algorithm realizes the high-precision prediction of poultry house temperature in different seasons and different climatic conditions, and the prediction accuracy is improved by about 23.7%-37.8% compared with other prediction algorithms.