Robust Fast Independent Component Analysis Applied to Mineral Resources Prediction
Mineral resources are one of the most important factors related to society development. Traditional methods of mineral resources prediction have some limitations and cannot satisfy the complexity of the prospecting geochemistry data. In this paper, Independent Component Analysis (ICA) is firstly applied to mineral resources prediction. FastICA, one kind of ICA algorithm, is applied to analyze geochemistry data, which is collected in gold deposit area of Inner Mongolia in China. The geochemistry data have 10855 samples and 18 elements, our attention is five elements , Ag, Au, Cu, Pb, Zn and their poly metallic-deposits. Before doing fastICA, we preprocess the observed data to satisfy the assumption of the data model which is that the mean of the data is zero and make the algorithm convergent faster. To enhance the robustness of ICA, in preprocessing step the outliers are modified in a certain range. Experiments show that fastICA outperforms PCA on prediction accuracy. FastICA increases reliability of results.
Xianchuan Yu Shihua Liu Jiamian Ren Ting Zhang
College of Information Sci. & Tech., Beijing Normal University, Beijing 100875, China No.722 Geological Group of Guangdong Bureau of Geology and Mineral Resources, Santou 515021, China
国际会议
The 12th Conference of the International Association for Mathematical Geology(第12届国际数学地质大会)
北京
英文
94-97
2007-08-26（万方平台首次上网日期，不代表论文的发表时间）