A Pilot Evaluation of Adopting the Apriori Algorithm Using WEKA Tool for Predicting Anemia and Blood Disorders Based on Laboratory Indicators from South Libya

Abstract

One of the most popular data mining algorithms for association rule mining is the Apriori algorithm. However, compared to Anemia data analysis are still restricted. Using the open-source WEKA tool, this study attempts to assess how well the Apriori algorithm extracts predictive association rules from Anemia data indicators from South Libya. Analysis was done on a dataset comprising 397 records and 14 features, such as CBC indices and demographics. 183 (NN), 187 (MH), and 27 (MA) comprised the target groups. Because the digital biomarkers in the sample were continuous, a range expert had to preprocess the data using a discretionary categorical filter in order for the algorithm to function properly. Six experiments were run with various parameters to discover the best algorithm settings. The ideal algorithm settings were found to be 50 rules, 0.8 minimum confidence interval, 0.1 minimum support, and the use of a lift metric to rank the rules. Fifty association rules were identified. MA class had the greatest lift value of 11.76, while MH class had values ranging from 1.75 to 2.10. The retrieved criteria were highly consistent with current medical knowledge, and potentially useful novel patterns, which may aid in the early detection of macrocytic anemia. The study finds that WEKA's Apriori is an effective and interpretable tool for Anemia data, making it a dependable software core for assisting early clinical decision-making in healthcare facilities. The study advises to handle unusual classes, comparing different association methods, and incorporating association rules and systems in future clinical applications.