Analyzing rare earth mine distributions in mainland China: a machine learning approach with k-means clustering and SVM.

This study integrates map information projection methods with machine learning algorithms to analyze the distribution of rare earth mines in mainland China. The information obtained through the map information projection method includes the latitude and longitude of the deposits, deposit type labels, and deposit names. This approach helps to overcome challenges related to the sensitivity of geological information. The acquired information was organized into a simple dataset containing only latitude and longitude information and a complete dataset containing additional information. These datasets were used to simulate the early and later stages of the research project, respectively. The K-Means algorithm was applied to the simple dataset, and the results demonstrated good clustering performance through specific validation. The Support Vector Machine (SVM) algorithm was applied to the complete dataset, and the analysis showed excellent classification performance, with relevant metrics (Accuracy, Precision, Recall, F1 Score) all around 90%. The experiments demonstrate that K-Means and SVM are suitable for information analysis in earth sciences and that they complement each other in research projects, being particularly applicable to the early and later stages of the project, respectively.The findings contribute to a more nuanced understanding of rare earth mineral distributions and underscore the potential for machine learning techniques to revolutionize geological sciences. [ABSTRACT FROM AUTHOR]