Building settlement forecast based on BP neural network

In order to obtain the law of the building settlement and forecast it effectively, neural network model was established for building settlement forecasting based on measured data, and an engineering example is shown to test and verify. Firstly, data of building settlement measured were normalized; embedding dimension was selected to establish the leaning samples. Mean square error (MSE) and mean absolute percentage error (MAPE) were used to evaluate the accuracy of the model. BP neural network forecasting model was established with example of the small high rise of China University of Mining and Technology (CUMT). The results show that MSE of 4#3 point is 2mm, and MAPE is 4.8%; the MSE of 8#3 is 3mm, and the MAPE is 3%. Both forecasting results are accurately and reliability which can meet the requirement of on-site engineering. The research provides a new approach of the building settlement forecast.