Applying Decision Tree Classification Techniques for Diagnose the Disease in Cow on Mobile Phone

Abstract

The purposes of this research are 1) to develop a model to diagnose diseases in cows using decision tree classification techniques 2) to develop applications for diagnosing diseases in cows on mobile phone and 3) to evaluate user satisfaction with this applications. This study collected data on factors associated with diagnosing cows from cattle raisers and specialists in the Phitsanulok province. The diagnostic models were created using three decision tree algorithms for performance comparison. The three algorithms consisted of J48, RandomTree and REPTree. The performance measured using cross-validation to evaluate the best diagnostic model. Random Tree algorithm was the best diagnostic model. The performance of RandomTree algorithm showed the accuracy of 99.47%.  The Root Mean Squared Error was 0.020, Precisionwas equal to 0.995, Recall was equal to 0.995 and F-measure was equal to 0.995. Therefore, this model was used to develop the application for diagnosing cow diseases  on mobile phone. This application showed details of disease, causes, symptoms and prevention of disease in cows. The user evaluation of the application  with 35 users who were cow raisers and general users  revealed high satisfaction. The overall average score was 4.01 and standard deviation was 0.55. Therefore, it can be concluded from the evaluation results  that the application is a effective application.