• OMOFUOMA, LUCKY OGHENOVO
• Okoli , Cecilia Nchedo

This study focused on comprehensive evaluation of four criteria for discriminant analysis procedure which includes Fisher, Welch, Bhattacharyya and Mahalalanobis distance. The objectives of the study were to evaluate the performance of each criterion based on classification accuracy, to ascertain the best fitted discriminant criteria, to evaluate the performance of each criterion based on addition of outlier to the original dataset and ascertain the consistency performance of each criterion based on simulated data for large sample sizes. The R- software package is employed for the data analysis. The analysis was conducted on original dataset (n=50) and on subsequent data. The null hypothesis assumption of homogeneity of covariance matrice was equally satisfied. The result of the study revealed that Fisher, Welch, Bhattacharyya criteria showed 86.7 percent (86.7%) accuracy rate. The finding also showed that in outlier detection that these criteria outperformed others with an accuracy rate of 87.5 percent (87.5%) and Mahalanobis demonstrated more efficient and best fit model having AIC value of 70.69 and BIC value of 91.37. Further findings from simulated data for various sample sizes revealed that Mahalanobis is consistently better than other criteria under this study. The study also highlighted the importance of considering trade- offs between computational complexity and interpretability when selecting criteria, acknowledging potential variability on dataset and study context. This work will be useful to researchers and practitioners to select the most appropriate criteria for their specific datasets and research objectives.

Keywords: Linear Discriminant Analysis, Fisher, Welch, Mahalanobis Distance, Bhattaryya Distance , Classification Accuracy.

"EXPLORING THE EFFECTIVENESS OF FOUR CRITERIA FOR DISCRIMINANT ANALYSIS: A COMPARATIVE EVALUATION", International Journal of Emerging Technologies and Innovative Research (www.jetir.org), ISSN:2349-5162, Vol.12, Issue 5, page no.j777-j790, May-2025, Available :http://www.jetir.org/papers/JETIR2505A70.pdf

"EXPLORING THE EFFECTIVENESS OF FOUR CRITERIA FOR DISCRIMINANT ANALYSIS: A COMPARATIVE EVALUATION", International Journal of Emerging Technologies and Innovative Research (www.jetir.org | UGC and issn Approved), ISSN:2349-5162, Vol.12, Issue 5, page no. ppj777-j790, May-2025, Available at : http://www.jetir.org/papers/JETIR2505A70.pdf