Integrated Data Mining Strategy for Effective Metabolomic Data Analysis

Disease diagnosis using molecular profiles has gained more attention during the last decades. Among the molecular diagnosis study, metabolomics has been a recently emerging field as promising tools for early detection of diseases. However, due to complexity and largeness of the metabolic profile data, data mining techniques have been essential to handle, process, and analyze the data, and also it is not obvious to apply the data mining techniques to such data, accordingly suggesting the need for suitable data mining strategies for the metabolomic data analysis. In this study, we propose required data mining procedures for effective metabolomics studies including description of current limit or future prospect and our approaches, consisting of preprocessing, dimension reduction, feature analysis and selection, classification, and automated data processing software.