| 定量成矿预测的人工神经网络模型 |
| |
| 引用本文： | 吕新彪,赵鹏大.定量成矿预测的人工神经网络模型[J].地球科学,1998,23(6):619-623. |
| |
| 作者姓名： | 吕新彪 赵鹏大 |
| |
| 作者单位： | 中国地质大学资源学院 |
| |
| 基金项目： | 国土资源部“矿产资源定量预测与勘查评价开放实验室”1997年基金 |
| |
| 摘 要： | 重点探讨了资源量预测人工神经网络ＢＰ模型（ＢＰ－ＭＲＰ）的建造和预测问题，特别对模型输入层（地质变量）、隐含层和输出层的构置和优选，模型的学习、检验和预测评价等问题作了较深入的讨论，尝试性地提出了模型输入变量的优化方法和模型检验准确率的计算方法，并以长江中下游地区为实例，给出了ＢＰ－ＭＲＰ模型的实际建造步骤和预测。结果证明，人工神经网络模型不仅能够模拟成矿地质因素和矿床特征（值）之间的非线性关系，
|
| 关 键 词： | 人工神经网络 多金属矿床 成矿预测 定量预测 |
MODEL OF ARTIFICIAL NEURAL NETWORKS FOR QUANTITATIVE PREDICTION OF MINERALS |
| |
| Abstract: | It mainly discusses how to make BP_MRP model, giving a detailed research on the designing and optimum selecting of the import, hidden and output layers in the model, and how to use the model to complete minerals prediction, including the study, acceptance inspection and assessment of the model. It also tries to give the methods which can select the assemblage of optimum varieties in the import layer of model and calculate the accuracy rate of prediction of the model. The steps and processing of construction and application of BP_MRP model are shown with an example, the middle_lower Yangtze River. All studies above_mentioned demonstrate that the ANN model not only simulated the nonlinear relations between the geological factors of ore_forming and the values of ore deposits, but also predicted the types of ore_forming environment, the kinds, number, size and richness of ore deposit simultaneously. |
| |
| Keywords: | artificial neural networks mathematical model of ore deposit quantitative prediction of minerals Fe_Cu_Au deposits middle_lower Yangtze River |
| 本文献已被 CNKI 维普 万方数据 等数据库收录！ |
| 点击此处可从《地球科学》浏览原始摘要信息 |
|
点击此处可从《地球科学》下载全文 |
|
