An improved fuzzy rule-based system using evidential reasoning and subtractive clustering for environmental investment prediction

Environmental investment prediction has attracted much attention in the last few years. However, there are still great challenges modeling, e.g., 1) effective environmental indicators must be accurately selected to avoid curse of dimensionality; 2) data reasonably downsize scale historical data; 3) higher interpretability and lower complexity models considered. To address above three challenges, a new model using fuzzy rule-based system (FRBS), evidential reasoning (ER) approach, subtractive clustering (SC) algorithm is proposed present work, called FRBS-ERSC. In this model, FRBS core component for modeling therefore provides good managers. Meanwhile, ER approach used as an improvement technique combine strengths different feature selection methods better indicator selection, SC another select data. An empirical case studied based on 31 provinces China ranged from 2005 2018. The experimental results show that FRBS-ERSC not only interpretable scalable but also produces satisfactory accuracy compared some existing models.