Development of knowledge based model for the diagnosis of sorghum diseases using Rule-Base approach

Abstract

The nuances of semantics, diction, and intent are too broad for a computer to understand directly. Instead, humans give the computer a set of rules which define the way to process incoming information. Hence, expertise in this particular field is apparently essential to help diagnose crop pest and disease whilst recommending a proper treatment or control procedure to farmers affected by the incident. Damage occurs, both on the field during the cultivation process and warehouse storage. These conditions will significantly affect the income of farmers and the world’s food supply. The aim of this research is to develop a knowledge-based model for prediction and prescription of sorghum diseases using the concept of Machine learning techniques (K- nearest neighbor, KNN) algorithm. The specific objectives include: To seek, create and preprocess dataset, To develop the front end of the system using Java programming language and the back end using PHP, To Develop and implement the K-nearest Neighbour algorithm, Train and test the Model, To validate the Model using Confusion matrix in order to determine the prediction accuracy, The software will bring an end to the manual method of application which leads to applying the wrong insecticide, fungicide etc. to the wrong crop and in a wrong quantity, It will be beneficial to the farmers in Adamawa State and their experts. Plant diseases are the most important reason that leads to the destruction of plants and crops. Detecting that disease at early stages will enable us to overcome and treat them appropriately. The process requires an expert to identify the disease, describe the method of treatment and protection. Identifying the treatment accurately depends on the method that is used in diagnosing the diseases. An expert system was developed using two different methods of plant diagnosis: Step by step description and graphical representational methods. The simulation was carried out using JAVA Net beans for windows.