Forecasting Study On Peasants' Income Of Shaanxi Province Based On BP Neural Network

The peasants' income is the core problem of rural economy. It is related to the social stability of rural area and the health development of national economy. The peasants' income forecast to future age is a new problem. It has important guiding significance and real value to workout a relevance supporting agriculture policy which increase the peasants' income. Especially the"three rurals"is commonly paid close attention now. How to improve the peasants' income catches everybody's attention on the modern time mingling to be deepened by gradually in city and countryside. Therefore, the thesis builds a BP Neural Network model to forecast the peasants' income from a brand-new research visual angle.Firstly, this thesis comments the background, purpose and significance of the research. It simply introduces a reality example that Artificial Neural Networks method applies to the forecasting field. The thesis sums up the current situation studying of home and abroad, expounds the thought and method of the research, and points out the innovative place.Secondly, the thesis introduces the history and basal principle of Artificial Neural Networks. It analyses and discusses the characteristics and application fields of Artificial Neural Networks by expounding the basic knowledge and fundamental model of Artificial Neural Networks.Thirdly, the thesis discusses the concept and effect of the peasants' income forecast, introduces the existing available forecast method, analyses several kinds of available models forecasting the peasants' income, and points out the defects of available models.Then, the thesis discusses the forecast model's establishment of BP Neural Networks based on peasants' income in detail. It mainly includes the following several aspects. First, it suggests a forecasts research step and forecasts feasibility based on the principle of BP Neural Networks applying to forecasts. Second, it discusses the key technology building forecast model based on BP Neural Networks, including the choice and pretreatment of sample book, the choice of input and output variables, the ascertainment of hide tier of node numbers, the choice of initial right and the threshold value, the choice of a active function, a training algorithm and parametric. It builds the rational network model finally. Finally, the thesis takes the demonstration analyses as background, analyses the data source of peasants' income forecast model in detail, takes the relevance data tape of peasants' income over the years as example sample book, builds a forecast model owing to neural networks drawing support from MATLAB neural networks tool case, compares and analyses forecast values and actual values. It testifies that the forecast method reveales the relationship between the peasants' income and the influencing factor within certain error range and can be used for forecasting future peasants' income by verifying the peasants' income forecasting example.