An optimized enhanced Intuitionistic fuzzy cognitive maps for groundnut yield prediction

Over the past decades, several types of crop yield prediction systems using different kinds of data mining algorithms have been developed in agriculture that supports cultivators to analyze the yield productivity. Among those techniques, Fuzzy Cognitive Map (FCM) based crop yield prediction has better efficiency, flexibility and ability to predict yield productivity. However, the performance of FCM was degraded due to some missing input data. Hence in this article, Intuitionistic Fuzzy Cognitive Map (IFCM) is initially used to improve the groundnut yield prediction with the aid of weather and soil parameters. The IFCM is built by considering the expert’s hesitancy in the computation of the causal relations between the concepts of a groundnut yield. On the other hand, the learning rate and stability of the IFCM are less due to fixed parameter based weight adaptation. As a result, a supervised multistep learning using the gradient method is proposed for enhancing weight adaptation of IFCM. The enhanced IFCM (EIFCM) estimate the current value of the weight matrix elements from the previous estimation history. Moreover, the learning parameters of the gradient method utilized in EIFCM are optimized by using Self-Organizing Migration Algorithm (SOMA) to reduce the iteration of the weight update. The experimental results prove the efficiency of the proposed OEIFCM in crop yield prediction in terms of accuracy, precision and recall.