Study On The Comprehensive Unit Price Prediction Based On Hadoop Platform

Bill of quantity valuation is widely carried out in the project field by the market leading price valuation method, determine the comprehensive unit price is the core content of the projectbill of quantity valuation. However, the current commonly used to determine the comprehensive unit price method or apply the quota price promulgated by the government to set prices, and then according to the market price of price adjustment, and calculate the corresponding comprehensive unit price, and did not fully realize the market pricing model, therefore, to comprehensive unit price reasonable determine research has important theoretical meaning and practical value. The main work of this paper is as follows:(1) Research on the valuation process, the composition elements of the comprehensive unit price and the calculation method of the project quantity list valuation mode, and determine the main elements of the comprehensive unit price prediction.(2) Research on time series autoregressive moving average prediction model algorithm, improved design for in nonlinear time series residual Auto-regression Prediction Model, the affecting factors of comprehensive unit price prediction, write corresponding prediction steps and flow chart, and the forecasting model of simulation program with the matlab program. Finally, combined with specific examples of data, verify the prediction model and the feasibility of the matlab program.(3) The cloud computing Hadoop platform based architecture and key technology of distributed file storage HDFS and distributed parallel programming model MapReduce processes running on a study, based on cloud computing Hadoop platform prediction of comprehensive unit price in parallel prediction model is established.On the computer installed cloud computing platform Hadoop, configure environment variables and operation configuration by Matlab’s built-in MapReduce parallel computing architecture to write map and reduce functions, and use the Matlab compiler calls to the Matlab prediction program, and finally the use of engineering examples of comprehensive unit price to predict and verify the feasibility in the cloud computing platform Hadoop for data processing.This study shows that the residual autoregressive prediction algorithms are studied, using Matlab for the realization of the program and its call to the cloud computing Hadoop platform of comprehensive unit price forecasting can reflect market price fluctuations timely and provide a new method for reasonable engineering price.