Learning under uncertainty for interpreting the pattern of volcanic eruptions

The overall goal of the research presented in this paper is to design an intelligent system to aid geologists in processing complex rock characteristics for interpreting eruption patterns, and thereby to aid eruption forecasting for volcanic chains and fields. The objective of this paper is to introduce a belief-based partially supervised classification method designed to deal with high uncertainty of geological data. A case study developed to show the feasibility of the presented method for correlation of tephra layers based on geochemical characteristics is also described. This method is not specific to geological data and can be used in other applications.