As a calculation model with high non-linear mapping ability, BP neural network has excellent abilities of non-linear approximation and high fault tolerance which make BP neural network useful in mineral resources assessment field. In the last few years, researchers have applied BP neural model in the assessment of targets and acquired some achievements. However, these applications did not analyze theatrical deficiency of neural network model and its effect on the results. This paper first applies the data of geology, geophysical prospecting, chemistry and remote sensing in 1 to 200000 scales to assess the targets of Cu-Ni sulfide deposit in the East Tianshan area. Secondly, the paper analyzes the two results sorted out by BP neural model based on different two groups of samples. Thirdly, it classifies the same data using the program based on the method of characteristic analysis. Finally, it compares and analyzes the two results and summarizes the key steps of mineral resources assessment based on BP neural network and verifies the classified result of BP neural network.%在深入研究BP神经网络算法的基础上,开发BP神经网络矿产资源评价功能模块并以东天山地区铜镍硫化物矿床为例开展成矿远景区预测工作.使用MRAS软件中的特征分析功能基于相同样本数据开展预测评价,并对两种模型产生的结果进行对比分析和交叉验证.证明BP神经网络程序的正确性,同时总结BP神经网络矿产资源评价应用的关键技术要点.
展开▼