A Soft Computing Approach to Rainfall Intensity Classification Using TRMM/TMI Data

Soft computing is tolerant of imprecision, uncertainty, partial truth, and is used to cope with complicated problems. In this paper, we take advantages of various soft computing techniques, fuzzy systems, neural networks, and evolutionary computations, to construct an ensemble classification model, the Enhanced Adaptive Hierarchical Classifier (EAHC). The EAHC utilizes fuzzy systems to gain extra data attributes and hence is able to construct classification models with higher classification accuracy. Furthermore, we integrate several data mining technologies, such as, decision tree, self-organizing map, and fuzzy IF-THEN rules which are used to build basic classifiers of EAHC. The integration also leads to EAHC achieving a better classification results. The EAHC has been applied to solve a critical, real-world problem, namely rainfall intensity classification. EAHC can achieve various goals through setting different fitness functions. Experimental results show that the proposed model is able to achieve high accuracy for rainfall intensity retrieval and outperforms previously published methods. Finally, we apply EAHC to a real typhoon case, which demonstrates relatively high agreement between GPROF and our algorithm.