Bidirectional Associative Memory Classifier for Personalized Recommendations for Farmers

This paper suggests Bidirectional Associative Memory based classifier for agricultural recommendation system. Use of ICT in agriculture sector enables farmers getting the right information at right time. Online web application for farmers can help them in getting information about crops, seeds, fertilizers, price information, their competitors and potential markets. Yet, not all farmers are equally educated to use online information since they have diverse economic and educational background. Hence the web application intended for agriculture sector must be personalized to suit the requirements of individual farmer. This research paper presents an approach for reducing the knowledge gap between farmers and agriculture experts. In this paper a framework for an agricultural website is proposed. The framework consists of different segments. Each segment is designed to fulfill the requirement of a group of users classified on the basis of their personal profile and information needs. Here we propose a novel dynamic classification algorithm based on Bidirectional Associative Memory that stores the profiles of farmers and their associated ratings for information requirement. When a user logs in, his/her profile is matched against the stored results in the model. The user is then redirected to an appropriate segment of the website. Finally, the recommendations are made by the website and made available to the user. A survey is conducted to generate the dataset and results show that proposed model has performed satisfactorily and classifies the users with reasonable level of accuracy.