Rainfall prediction in the northeast region of Thailand using Modular Fuzzy Inference System

Abstract

Abstract-In water management systems, accurate rainfall forecasting is indispensable for operation and management of reservoir, and flooding prevention because it can provide an extension of lead-time of the flow forecasting. In general, time series prediction has been widely applied to predict rainfall data. The conventional time series prediction models or artificial neural networks can be used to perform this task. However, such models are difficult to interpret by human analyst. From a hydrologist's point of view, the accuracy of the prediction and understanding the prediction model are equally important. This study proposes the use of a Modular Fuzzy Inference System (Mod FIS) to predict monthly rainfall data in the northeast region of Thailand. The experimental results show that the proposed model can be a good alternative method to provide both accurate results and human-understandable prediction mechanism.

A modular technique for monthly rainfall time series prediction

“…This study proposes a use of modular technique to perform monthly rainfall time series prediction. This study could be seen as an improvement of the work reported in [2]. …”

“…To evaluate the prediction accuracy, the proposed model was compared to hydrological common-used prediction models, namely, Autoregressive Moving Average (ARMA), Artificial Neural Network (ANN) [2], [3], [4], [5] as well as Fuzzy Inference System (FIS) [19], [20]. Furthermore, the proposed model is also compared to the model without aggregation method [2].…”

“…Therefore, two antecedent rainfalls are considered as input for each sub-module. This 2-lag input has been proved in the work of [3] and [14] that it has sufficient information for modular model to provide accurate results.…”

“…However, the prediction mechanisms of those models are in parametric form, which are difficult for human analysts to analyze [2], [3]. This could be viewed as a disadvantage of those models.…”

“…However, those models are difficult to understand due to the black-box nature of the prediction mechanism. Kajornrit et al [2], [3] proposed the use of modular fuzzy inference system to make the prediction model more transparent and easily interpretable by human analysts. However, such models still lack of the capability to address the uncertainty in the time dimension.…”