Systematization of Jamu using data-intensive science and machine learning approaches
Summary:
Jamu is a popular traditional medicine from Indonesia. The formulation of Jamu is composed of a single plant or a mixture of several plants based on users’ experience for decades or even hundreds of years. Therefore, systematization of the formulation of Jamu is needed to understand Jamu better, develop basic scientific principles of Jamu, and meet the Indonesian Healthcare System requirement. Many approaches can be utilized to build scientific background about Jamu, including data-intensive science and artificial intelligence.
This study is intended to explore and identify interesting patterns in the formulation of Indonesian Jamu by applying data-intensive science and artificial intelligence, especially machine learning methods. Initially, we introduced a new method to predict the relations between plant and disease using network analysis and supervised clustering. The predicted plant-disease relations were evaluated in the context of previously published results and produced excellent predictions. In addition, the analysis was extended by including metabolites information of the plants used as Jamu ingredients for predicting Jamu efficacy and identifying essential metabolites. The Support Vector Machine with linear kernel and Random Forest obtained good classification models if we combined these classifiers with Single Filtering algorithm and Regularized Random Forest. We also identified 94 significant metabolites associated with twelve efficacy groups by applying the inTrees framework.