| Agricultural mechanization was the main way to improve agricultural productivity,optimize the industrial structure of agriculture,promote the transfer of rural labor force,enhance the efficiency of rural land and reduce the labor intensity of farmers.The level of agricultural mechanization was one of the important indexes to measure the development level of agricultural mechanization in a certain area.Since the promulgation of the "Agricultural Mechanization Promotion Law" in 2004,the development environment of agricultural mechanization in China has been significantly optimized,the industrial structure has been constantly adjusted,and the level of agricultural mechanization has been continuously improved.In 2016,the operation level of agricultural mechanization in China reached 65.2%.The main grain crops production mechanization was advancing rapidly,and the main economic crops production mechanization achieved a significant breakthrough.Although the development of agricultural mechanization in China has achieved excellent results,there were still many problems in the development of agricultural mechanization in China compared with the developed countries.The mechanization of agriculture faced many challenges.The level of agricultural mechanization was still at the stage of intermediate development.Therefore,it was of great significance to correctly understand the level of agricultural mechanization and to realize its accurate prediction for the formulation of relevant policies and the rational allocation of resources.This paper analyzed the influencing factors of the level of agricultural mechanization operation comprehensively.Three different prediction methods were adopted to predict the operation level of agricultural mechanization in China,and the suggestions for the development of agricultural mechanization were put forward.From the four aspects of social and economic development,production factor input,output benefit,service and guarantee,fifteen main influencing factors were systematically extracted,and the data of each index were collected.The factors of economic and social development including four second level indexes,respectively,were a production of GDP and agricultural value,per capita net income of rural people's three major grain crops in the average price of standard work.The factors of input of production factors included six second level indexes,namely the total power of agricultural machinery,the total input of agricultural machinery,the number of rural employment personnel,the number of rural farm workers,the number of farm machinery households,and the amount of agricultural diesel fuel.The output benefit factors included three second grade indexes,which were the average number of three grain crops,the output of grain per unit area,and the total income of agricultural machinery use.The service and security factors included two second level indexes,which were the number of farm machinery repair outlets and the number of agricultural machinery safety supervision institutions.Fifteen factors were analyzed by principal component analysis,and two main components were extracted from them,and 96.445% of the original information was retained.The simplification of the data structure of the influencing factors was realized on the premise of reducing the information loss.This study established a wavelet–BP neural network prediction model by targeting the nonlinearity and non-stationary features of the data under the fundamental principle of wavelet analysis and BP neural network.First,the major factors that influence the operation level of agricultural mechanization were determined and analyzed,and dimensionality was reduced through a principal component analysis.Second,the time series of the operation level of agricultural mechanization and the principal component series of the influencing factors were decomposed to obtain low-frequency and high-frequency components.A BP neural network prediction model was built for the low-and high-frequency components.Lastly,the obtained low-frequency and high-frequency components were examined through linear super position,and the final prediction results are obtained.The proposed method is verified by predicting the operation level of agricultural mechanization in China.Results showed that the wavelet–BP neural network prediction model could perform accurate prediction.The model evaluation indices,namely,average relative error,root-mean-square error,Theil IC,consistency indicator,effective coefficient,and excellence rate,were 0.44%,0.293,0.0024,0.90,0.9727,and 100%,respectively;these indices were superior to those of conventional and other models.To explore the effects of different combination forecasting models on the operation level of agricultural mechanization,the time series of the operation level of agricultural mechanization and the historical data from 2001–2012 were selected as the research object and training sample,respectively,in this study.The exponential curve method,the cubic exponential smoothing method,and the gray forecasting method were adopted to construct the single forecasting model.Moreover,the least sum of squared error(LSSE),Shapley,and IOWGA operators were selected to construct the combination forecasting model and forecast the operation level of agricultural mechanization in 2013–2015 based on the forecasting results of the single model.The comparison results showed that the forecasting effect of the IOWGA combination model was the best,followed by that of the combination forecasting model based on LSSE,whereas the Shapley combination model exhibits the lowest forecasting accuracy.The IOWGA combination forecasting model could collect valid information from each forecasting model and assign different weights according to the level of forecasting precision,while exhibiting better forecasting effect and stability.Its relative error could be controlled at 1%.The IOWGA combination forecasting model exhibited the advantage of forecasting the operation level of agricultural mechanization in China.The time series of the operation level of agricultural mechanization was selected as the research object,according to the characteristics of nonlinearity and randomness of data samples,the advantages and disadvantages of grey model and Elman neural network model were analyzed of constructing the grey Elman neural network model based on genetic algorithm optimization.Moreover,the gray genetic Elman neural network model was adopted to fit the data of 2000-2013 and to predict the operation level of agricultural mechanization in 2014-2015.Results showed that the gray genetic Elman neural network model could exhibit better prediction effect and its average relative error was 0.18%.This model fully played the advantage of the gray model to weaken randomness and Elman neural network to deal with the nonlinear problem,which was suitable for the prediction of the operation level of agricultural mechanization operation.Considering the prediction results,the corresponding suggestions were put forward from the following four aspects,which were strengthening promotion planning,realizing the sharing of agricultural machinery,implementation of land scale management,adjusting grain subsidy system.Through the above research work,it will provide some references for the development planning of agricultural machinery in China,and provide theoretical basis for the formulation of relevant policies and regulations for agricultural machinery,so as to enhance the rapid and sound development of agricultural modernization in China. |
PDF Full Text Request