基于主成分分析的Fisher判别分析法开展杂卤石识别研究

Based on the Principal Component Analysis and Fisher Discriminant Analysis to Identify the Polyhalite

摘要: 四川盆地下中三叠系海相钾盐广泛分布，主要发育的固态钾盐——杂卤石多与石盐、硬石膏、岩盐等矿物共生，岩性十分复杂。因此，准确识别出杂卤石就显得十分重要。笔者以测井方法理论和数学统计理论为基础，结合主成分分析方法，建立fisher判别模型开展研究区岩石矿物定性识别研究，整体的判别正确率达到了82.9%，效果明显。与传统测井识别方法——交会图法相比具有识别准确率高、操作简单、识别速度较快等特点，值得进行推广使用。

Abstract: The potassium salt of the lower Triassic in the Sichuan Basin is widely distributed. The main de-velopment of polyhalite is associated with minerals such as rock salt, anhydrite and rock salt, and the lithology is very complicated. Therefore, it is very important to accurately identify the polyhalite in the Sichuan Basin. Based on well logging method theory and mathematical statistics theory, the author combined the principal component analysis method to establish a fisher discriminant model to carry out qualitative identification of rock minerals in the study area. The overall discriminant accuracy rate reached 82.9%. Compared with the traditional logging identification method—cross plotting method, it has the characteristics of high recognition accuracy, simple operation, and high recognition speed. So, it is worthy of promotion and using.