High-rise Residence Project Costs Estimation Based On The Principal Component Analysis And BP Neural Network

The project cost which is estimated quickly and accurately is the focus of many scholars and insiders. There are many factors that affect the construction cost. If we consider all the factors, the workload of the cost increased exponentially. Based on past experience, if only select representative indicators, will not convincing enough because they don’t have the objectivity. In this article based the principal component analysis, a project costs estimation module was established. This model can made a number of factors to become a minority controlling factors, through this few factors to complete the estimate.Firstly, per-processing technique of cost estimation methods at home and abroad were compared and analyzed, the paper studied its rationality and inadequate, on the basis of the introduction of the principal component analysis. Principal component analysis by linear transformation of the original data into another set of mutually independent new variables, these new variables are mutually independent. Principal component analysis through the raw data standardized to make the data analysis easily and eliminate the influence of the data dimension and magnitude. The paper in order to make the principal component analysis is more suitable for non-linear processing of the data analysis, the value of the feature extraction, linear dimension reduction and other characteristics to determine the direction of the vector were improved, which reducing the loss difference information on the degree of variation of each index and improved extractability eigenvalues.Finally, using the results of the above theoretical analysis, BP neural network algorithm is proposed based on the project cost estimation method based on principal component analysis, given the specific workflow. The established model is examined with the high-rise residential project cost data, and the results show that the predictive model so established demonstrates ideal accuracy with accepted principal component analysis, and it has some practical significance.