Study On The Cost Estimation Of Residential Projects

With steady growth of the real estate market in China, the market competition of the real estate will become more intense in the future. The seller’s market will gradually transit to a buyer’s market, housing price and the profit of enterprise turning rational.The investment decision-making of construction project has gradually entered the stage of fine management. The quota government issued usually takes a relatively long period with a lag.resulting in the inaccuracy of the cost estimation,far behind the current price level.The quick and accurate cost estimation with the characteristics of the residential house will not only save the time cost on biding, but also help to make a good finiacing plan. This way, this paper aims at making a quick and accurate cost estimation assisting the enterprise in decision-making. The main work of this paper has the following four points:1.This paper comes up with a question,make a research on the background and the significance of the cost estimation, review domestic and foreign literature on cost estimation, make a research on the theory of cost estimation on construction project and the case-based reasoning.2.This paper introduces a model based on the theory of case based reasoning and cost estimation.Firstly establish the database of residential house with data preprocessing, secondly use the gray correlation calculation to retrieval case, thirdly calculate the ratio of similarity and the summary of similarities as the weight of target project, fourthly calculate the weighted summary of historical cases to solve the case, finally, analysis the accuracy of case-based reasoning estimation model.3.This paper introduces two model based on the theory of support vector regression machine.The first one establish the database of residential house without filter processing, seeking the optimal combination of parameters,train the SVR model,estimate the cost of projects and analysis the accuracy of the SVR model.The second one establish the database of residential house with filter processing through the calculation of gray correlation, seeking the optimal combination of parameters,train the SVR model,estimate the cost of projects and analysis the accuracy of the SVR model.4. This paper introduces two model based on the theory of support vector regression machine and particle swarm optimization algorithm. The first one establish the database of residential house without filter processing, optimize the parameters through pso algorithm. using mean square error of support vector regression machines as fitness function, train the model, estimate the cost of projects and analysis the accuracy of the PSO-SVR model. The second one establish the database of residential house with filter processing through the calculation of gray correlation, seeking the optimal combination of parameters through pso algorithm, using mean square error of support vector regression machines as fitness function, train the model, estimate the cost of projects and analysis the accuracy of the PSO-SVR model.5. Make comparisons of the five models above.The five models introduced above all have small relative error within 10 percent, consistent with the policies and regulations during the decision-making phase of project, offering a little reference for initial investment decision.