Cost Prediction Of Market-Oriented Reform Of Ordinary Highway Maintenance Based On Improved Support Vector Machine Algorithm

As a public service,highway occupies a very important position in the economic development and people’s livelihood.By the end of 2014,the total mileage of highway in China is 4.4639 million kilometers,the total mileage of ordinary highway is 4.352 million kilometers,accounting for 97.5% of the total mileage,it can clearly be seen that ordinary highway plays an important role in the whole highway system.However,the development of traditional ordinary highway maintenance system did not keep up with the pace of construction scale,throughout the country,in most parts,it is still in place the ordinary highway management and maintenance system under the planned economy.Under this system,the state allocated directly maintenance funds to the various conservation authorities according to plan,management departments develop highway maintenance plan with conservation funds,not only cannot meet the actual amount of highway maintenance,but cannot stimulate the efficiency of workers.With the introduction of national policy of reforming institutions,many parts of the country’s ordinary highway maintenance industry has gradually taken a number of reform measures,it aims to push maintenance work to the market through reform,in order to solve the many ills in maintenance work.However,there is not much area and conservation management unit that really thoroughly complete market-oriented reforms,therefore,this article stand in the perspective of the government administration to discuss ordinary highway maintenance market reform cost prediction,it aims to provide a reference for the national ordinary highway maintenance market reform,while providing a reference for other domestic reforms.Firstly,the paper analyzed the related researches home and abroad,based on previous research results,this paper defines the meaning of ordinary highway maintenance market reform costs.Actual work in various stages of reform to be accomplished,analysis of the cost of reform which may arise,summary results of the analysis will be the cost of reform into direct costs,adaptation costs,frictional costs,risk costs,the loss of five categories resonance,and elaborated on the actual content of each cost included.Secondly,aim to the fact that reform costs is small sample,this paper proposed support vector machine to solve the problem of cost prediction.In view of theforecasting performance of the algorithm is affected by several key parameter,and currently it stays in a spreadsheet stage to determine the parameter values,which restricting the computing speed and prediction accuracy,to further improve the prediction performance of SVM,in this paper,the use of intelligent algorithms to improve the traditional SVM,and cost of the reform in the hope to obtain more accurate predictive value.Then,NAQ area is used as the actual case,it described in detail how to use the improved SVM to predict the cost of reform,predicted cost value of the reform has got in the region.On this basis,traditional SVM is used to calculate the predicted cost value of the area,the predicted values obtained by the two methods were compared,which shows that improved SVM can effectively improve the prediction accuracy.Finally,this paper link the cost of reform predicted value and the actual situation in the region NAQ,came up with the mode of financing,recommend gradual reform path and specific reform cost control measures,it explained how the predicted reform cost value of this paper use in the practical application of the reform process.