Study On Cost Prediction Of Prefabricated House Based On Improved Extreme Learning Machine

The prefabricated house is one of the main forms of residential industrialization.It has the advantages of fast construction speed,energy saving and environmental protection,and strong quality control,which is in line with the development concept of green building.At present,the state and various provinces and cities have introduced policies related to housing industrialization to promote the development of fabricated housing.However,the development of fabricated houses is not optimistic.The reason is that the high cost is the main factor limiting its development.Therefore,how to effectively control the cost of prefabricated houses is an unavoidable problem in the development process.In order to control the cost of fabricated houses,this paper starts from the factors affecting the cost of fabricated houses and predicts the cost of prefabricated houses.Mainly done the following work:(1)By analyzing the literature on fabricated houses,the necessity of researching the cost of fabricated houses is pointed out.The cost structure and characteristics of the prefabricated house are analyzed,and the cost of the fabricated house is studied.(2)Analyze the influencing factors of the cost of the assembled house,refer to the relevant literature and expert opinions,construct the initial cost forecasting index system,use the rough set importance theory and the analytic hierarchy process to calculate the weight of the cost forecasting indicators,and screen out the main cost forecasting indicators.reduce the complexity of the prediction model.(3)On the basis of analyzing the common cost prediction model,choose the ELM model with fast running speed and ability to process small samples as the modeling foundation,and optimize it with the PSO algorithm with strong global optimization ability.Through the training simulation of the collected sample data,the error curve between the predicted cost and the actual cost is drawn,and compared with the BP neural network and ELM model prediction results,the reliability of the prediction model is verified.The research of the thesis can effectively solve the cost management problems faced by the prefabricated houses,which is of great significance for promoting the industrialization of the houses.