prototype chatbot system for automatic answering questions on coconut diseases and insects with price forecasting

Abstract

The objectives of this research are: 1) to study and collect data on disease and insect incidence in coconut, 2) to develop a prototype chatbot system for automatic answering questions about coconut diseases and insects with price prediction, and 3) to compare the efficiency of three forecasting methods including the long-term constant average, single moving average, and single layer exponential smoothing. The experimental results show that the single-layer exponential smoothing method gives the smallest root mean square error with 1.24, followed by the single moving average method give the root mean square error with 1.27; and the last is the long-term fixed average method, the root mean square error equal to 20.60. Therefore, single-exponential smoothing method is used as a model for forecasting coconut prices in chatbots. The evaluation of the effectiveness of the prototype chatbot system for automatic answering questions on coconut diseases and insects with price forecasting by 5 experts received an average satisfaction score of 3.5, which is at a good level.