DESIGN AND IMPLEMENTATION OF AN INTEGRATED SYSTEM FOR AGRO-BASED INVENTORY MANAGEMENT SYSTEM AND CROP YIELD PREDICTION

Agriculture still remains the largest employer of labour and a major contributor to the economy of most developing countries. The proposed system is an integrated system for inventory management system (IMS) and crop yield prediction. An integrated approach combining inventory management and crop prediction as a single system improves farm management and better decision making among farmers and stakeholders in Agriculture. The proposed system uses Random Forest Model for predicting crop yield based on soil, average rainfall, temperature and historical yield in a given area while the IMS uses a graphical User Interface (GUI) for taking inventory and storing information in a relational database. Several literatures on inventory management and Machine learning techniques have been reviewed lending credence to the adoption of a database model and random forest algorithm for the development. An Object Oriented Analysis and Design Methodology and Agile methodology (for the predictive model) were adopted for the development. The proposed system was tested using various real-world entities for performance bench mark. Both the root-mean-square value and F1-score revealed the system performed better in terms of precision and accuracy. The result of this project is a well developed inventory management system capable of crop yield prediction with correlation coefficient of 0.5335, relative Absolute error of 71% indicating the system performs better than a naïve baseline.