Skip navigation
st. Mary's University Institutional Repository St. Mary's University Institutional Repository

Please use this identifier to cite or link to this item:
Title: Application of Data Mining with Knowledge Based System for Diagnosis and Treatment of Cattle Diseases: The case of International Livestock Research Institute (ILRI) Animal Health Center Addis Ababa
Authors: Fantahun, Zerihun
Keywords: - Cattle disease, Data mining, Knowledge based system, Rule based, Integrated with the Knowledge Base System,
Issue Date: Jul-2021
Abstract: Ethiopia is one among the nations that possesses the largest livestock population in the African continent with an estimated 56 Million of cattle, 58 Million of sheep and goats and 10 Million of equines, 1 Million of camels and 57 Million of chicken. Ethiopia has great potential for increasing livestock production, both for local use and export. However, development has been constrained by numerous reasons. In this study, the possibility of integrating data mining result with knowledge based system is realized and explored. The integration process begun by taking samples of ILRC dataset. The dataset is preprocessed and made suitable for mining steps. Due to several limitations in acquiring knowledge for knowledge base from domain experts in the area of diagnosis and treatment of cattle disease, integrated (manual and automated) knowledge acquisition techniques were used to acquire knowledge. Data mining has proven to induce hidden knowledge from large collections of datasets. Hence, data mining classifier, JRip is employed for knowledge acquisition step since it has performed best among the selected classifiers with an accuracy of 97.68%. To identify the best prediction model for diagnosis and treatment of cattle disease, 6 experiments for three classification algorithms, namely J48 pruned, Naïve Bayes and JRip under ten-fold Cross- Validation test option and percentage split test option were conducted. Finally, by conducting objective and subjective interestingness measure, the researcher decided to use rules that are generated by JRip classification algorithm model for further use in the development of knowledge base system because it registered better performance than J48 and Naïve Bayes with 97.68%, 96.65% and 95.42% evaluation result in 10-fold cross validation respectively. The prototype Knowledge Based System, which provides advice for Animal Health Workers about diagnosis and treatment of cattle disease was developed using SWI-Prolog 7.7.13 with NetBeans 8.2. The proposed Knowledge Based System has Knowledge Base, Inference Engines, Explanation Facility and User Interface. Then 70 test cases were prepared to evaluate the performance of the proposed system. Finally, system performance evaluation, testing and user acceptance testing were conducted. User acceptance testing is performed based on seven criteria of evaluation. Selected domain experts are trained and used the system to evaluate how much the KBS meets their requirements. The system on average scored 84.85% based on user acceptance evaluation.
URI: .
Appears in Collections:Master of computer science
Master of computer science
Master of computer science
Master of computer science

Files in This Item:
File Description SizeFormat 
Cover page.docx149.2 kBMicrosoft Word XMLView/Open
Zerihun Fantahun - Final Thesis.docx1.83 MBMicrosoft Word XMLView/Open
Show full item record

Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.