The Development of Decision-Making Models for Buying the Fattening Cattle Using Data Mining Techniques: A case study of the Buket-Chekha Beef Cattle Social Enterprise

Abstract

“The Buket-Chekha Beef Cattle Social Enterprise” is a group that operates its own business of food transformation of frozen beef products. Frequently, some farmers face financial loss due to a lack of experience in selecting appropriate cattle as input in the transformation process. Therefore, researchers have used the RapidMiner Studio to develop the 5 data mining models, including Decision Tree (C4.5), Iterative Dichotomiser 3 (ID3), Naive Bayes (NB), Random Forest (RF), and K-Nearest Neighbors (k-NN) models, for purchasing decisions of fattened cattle. The case study was based on the 416 purchase records of the Buket-Chekha Beef Cattle Social Enterprise from 2018 to 2022. The data cleaning and feature selection were pre-processed before building the models. Then, the data were separated into 10 folds to cross- validate between the train and test sets. As the decision accuracy measures the performance of the models, the experimental results show that the RF decision tree model has given the highest accuracy at 98. 53 % with 9 correlation rules, followed by the accuracy of C4.5, NB, ID3, and k-NN models at 97.02 %, 95.00 %, 94. 75 % , and 93. 54 %, respectively. Furthermore, the significant factors in the model were weight before slaughtering, body length, tooth structure, height, species, networking, fattening mean, fattening period, age, and purchase price.