Forecast of off-season longan supply using fuzzy support vector regression and fuzzy artificial neural network

•NN and SVR were discussed and proposed for forecasting of off-season longan supply.•Reducing NN and SVR proposed for improving computational runtime.•NN and SVR forecasting models were extended by fuzzy algorithms.•Totally, six proposed models were compared in term of their efficiency.•Multi regression was compared and over fitting was verified.

An over-supply crisis in longans in northern Thailand adversely affected farmer income. Cultivating longans off-season was adapted as an alternative solution to this over-supply problem. However, lacking information management and analysis, over supply occurred even during the off-season, leading to a slump in the sale price. Supply forecasting plays an important role in solving this problem. To solve this problem, we proposed a systematic approach for off-season longan forecasting using neural network, fuzzy neural network, support vector regression and Fuzzy Support Vector Regression (FSVR). In addition, grid search was applied to each support vector model to find its optimum architecture. Real data sets were used to evaluate and compare the effectiveness and efficiency of the algorithms. The experimental results showed that FSVR was the most effective forecasting technique.