Based on BP artificial neural network model, the prediction model is established. Taking the settlement detection data from phase 1 to phase 8 of a district building as input data and output data, the network model is trained. The paper also compares actual detection value to pre- diction value from phase 9 to phase 12. The results are relatively ideal, which testifies feasibility of applying BP artificial neural network model to predict building settlement.%以BP人工神经网络模型为基础，建立预测模型，以小区某栋建筑物l期～8期的沉降观测数据为输入数据和输出数据，对网络模型进行训练，并对9期～12期实际观测值与预测值进行了比较，结果比较理想，从而验证了采用BP人工神经网络模型进行建筑物沉降的预测是可行的。
展开▼
