Research On Engineering Cost Index Based On Data Mining

Facing the impact of the big data age,all kinds of complicated engineering data after collecting,sorting,cleaning,restructuring has great value,how to combine the method of data mining with these precious resources,is the realization of engineering cost industry sustainable development,implement the strategy of information management,accelerate the urgent needs of the service industry transition cost.In order to meet the requirement of high-speed development of socialist market economy,the construction project cost management must be changed completely rely on the government's quota valuation model.Under the premise that the government issued the quota and the regulation,the determination of the cost of the project should be determined by the market conditions and the production and management level of the enterprise.Therefore,in view of the engineering cost management department,this paper puts forward the reform research on the project cost index measurement system.Based on the theory of building engineering cost and the related theory of statistics,a complete cost index measurement system is designed under the list valuation mode,and the data mining theory is applied to the management of cost information.Including information screening,calculation,prediction and other methods.Finally,based on this research,we set up a set of information platform for building engineering cost index.In order to obtain the construction method applicable to the cost index system of the construction project in China's information construction,this paper systematically collates and analyzes the domestic and international literature on the calculation method and system setting of cost index.By collecting the eigenvalues of typical projects in recent years and investigating the setting of cost index in each city,a complete set of mathematical models of index calculation system,screening system and prediction system are established.This paper explains index theory,data mining theory,knowledge management theory,BP neural network,genetic algorithm and discusses its scope of application and value for this paper.This paper studies the setting and publishing methods of the index system in North America and the United Kingdom,and conducts research andanalysis on the six major cities in China.The results show that the index systems of different cities are quite different,mainly reflected in the following three aspects:First,the release form is different.Hong Kong,Chongqing,Shenzhen and Chengdu are the fixed base index,while Jiangxi adopts changed base Index.Tianjin has both a fixed base index and a changed base Index.Second,the issuing agencies are different.The setting and distribution of the Hong Kong cost index system is similar to that of the UK.The Architectural Services Department is responsible for publishing the government-level cost index,and the bid price index compiled by the two surveyors of Libby and Weining Xie also plays an important role.However,the construction cost index of the Chinese mainland is mainly compiled and published by local governments of various provinces and cities.Third,the scope of preparation is different.The Hong Kong Cost Index will compile a cost index and a price index.The Mainland China Cost Index will compile a single index and a composite index.This paper analyzes the construction project cost index calculation system at domestic and foreign.based on the characteristics of the current construction market,the compilation and design of the construction project cost index calculation system is determined by the list pricing model.To classify construction projects and propose to mainly compile three major categories of indices.they are the construction project cost index,the construction project expend index and the construction project consumption index.This paper has studied the sample screening methods of typical projects.When calculating the construction cost index,if the different types of project cost information are mixed together,the cost index will be lost,so it is very important to select typical construction projects as measurement samples,and it is necessary to summarize the similar projects after the division of the building functions and structural forms.Due to the scarcity of samples and the influencing factors during the construction period,in the process of screening typical projects,this paper collected100 construction project information,and adopted the "Expert Evaluation Method","Gray Connection Method","T Test Method" and "Decision Tree Method".Firstly,the "expert evaluation method" is used to select the eigenvalues of the sample feature values.Then use the "grey correlation method" to improve the conclusion of the "expert selection method",determine the eight eigenvalues,and then use the "Ttest" to screen,among the eight characteristic attributes,the three characteristic attributes of engineering location,layer height and structure type have no significant impact on the project cost,so they are eliminated.Finally,introduced the Decision Tree Method,the supplementary T Test Method can only test the dependent variable and cannot test the deficiency of the sample itself.Combine the four methods,from primary selection to perfection,and improve the accuracy of screening.This paper studies the measurement model of the cost index.The collection method of measurement information is mainly the manual collection method.A large number of senior information personnel in the industry are selected to perform statistics and report on the engineering information,and the engineering information through the audit can enter the system data.The collection method and release method of cost information were studied,and a complete information collection table was designed for the information collector.Adjusting the cost index from two aspectsFirst,in the case of the corresponding consumption of typical projects,the re-grouping of typical projects can obtain corresponding indices and indicators;Second,in China,due to the increasing demand for building materials in the market,the prices of major building materials such as steel,wood,cement,etc.have a great impact on the project cost.Therefore,this paper uses the adjustment formula to adjust the corresponding materials.This paper has studied the prediction of the construction project cost index.Due to the lack of eigenvalues of construction engineering samples and the small number of samples,the use of BP neural network in the prediction of future cost index has limitations,which will reduce the accuracy of prediction.Therefore,it is necessary to combine the characteristics of construction engineering samples to modify the original BP neural network.The project takes the high-rise residential and apartment(including commercial residential)projects(more than 12 floors)in the construction project of Wuhan as an example to carry out an empirical analysis of the construction project cost index forecast.The case selects the cost index of labor,materials and construction machinery that has a great influence on the construction cost,and analyzes and predicts the cost index,and collects data of various prices and construction price index from the third quarter of 2014 to the fourth quarter of 2016.information.Through calculation,using GA-BP neural network model based ongenetic algorithm optimization,the prediction accuracy of construction engineering cost index is about 97%,which has a key reference significance for construction project investment decision.Finally,this paper has studied the information collection system of the construction engineering cost index platform.The paper designed the information collection system.The platform users collected the core construction cost information of the construction project on the collection platform and established the regional engineering cost information database.Through the system data cleaning,auditing and processing services,the local construction cost station will regularly release the construction engineering index to the society to improve the efficiency and service level of cost management.