Nonlinear mathematical modeling in the model of village cultural industry

Abstract

Abstract To solve the problem that the current forecasting methods cannot describe the long-term correlation of traffic, which leads to low prediction accuracy, the author proposes a mathematical modeling, forecasting and analysis method for village nonlinear traffic. The original nonlinear traffic data collected from the village comes from each base station, the information contained is uneven, pre-processing the collected data, eliminating abnormal values and duplicate data, and supplementing the missing data. Nonlinear traffic contains limited information, so a random forest algorithm is used to extract traffic characteristics and reduce data processing dimensions. The nonlinear traffic characteristics of the village are convolved, and the cross entropy function is used as the loss function, the feature vector of the input prediction model is deeply learned, and the communication traffic prediction results are obtained. Taking the traffic data of the communication operation enterprise’s base station as the test data, the experiment results show that, in the test with 1 million pieces of data, the decision coefficient of the mathematical modeling, prediction and analysis method of village nonlinear flow designed by the author is 0.9599, which is 0.1267 and 0.1431 higher than the prediction and analysis method based on genetic algorithm and fuzzy clustering algorithm respectively. In the modeling and prediction of nonlinear flow, the determination coefficient of the method proposed by the author is closer to 1, the fitting degree of this method is better than that of the contrast method, and it is adaptive in the real scene with a large amount of data. It is proved that the mathematical modeling and prediction analysis method designed in this design can reduce NRMSE and MAPE, improve the determination coefficient of prediction results, and provide the basis for village analysis.