A New Discrimination Method for Classification of Multivariate Variables: Comparative Analysis

Okoro; Joyce Adaobi; Biu O. Emmanuel

doi:10.56201/ijasmt.vol.12.no7.2026pg34.55

📄 Download Paper

A New Discrimination Method for Classification of Multivariate Variables: Comparative Analysis

Okoro, Joyce Adaobi, and, Biu O. Emmanuel

Abstract

This study developed a New Discrimination Method for the classification of multivariate data and compared its performance with selected existing discrimination techniques. The variables used in the analysis were seven tailoring measurement features: length, shoulder length, sleeve length, neck, waist length, chest, and round sleeve. Data were collected from 200 randomly selected adult individuals (male and female). The discrimination methods examined in this study include Linear Discriminant Analysis , Quadratic Discriminant Analysis , Maximum Likelihood Discriminant Analysis , Nearest Neighbor (NN) classifier, and the proposed New Discrimination Method base on the median value. Data analysis was carried out using spreadsheet functions in Microsoft Excel. The proposed NDM was inspired by the nearest neighbor classification approach and utilizes the Euclidean distance function for assigning observations to predefined classes. Experiments were conducted on both small and large datasets to evaluate the effectiveness of the proposed method. The results indicated that the NDM achieved higher classification accuracy and improved computational efficiency compared with the existing discrimination methods. In addition, threshold values obtained for the different methods confirmed that the classification criteria adopted were adequate for reliable classification.

Keywords

Tailor MeasurementsClassification CriteriaThreshold ValueLinear Discriminant Analysis and Quadratic Discriminant AnalysisMaximum Likelihood Discriminant AnalysisNearest Neighbor classifier

References

Alsammak, I., Sahib, H., & Itwee, W. (2020). An enhanced performance of K-nearest neighbor (K-NN) classifier to meet new big data necessities. IOP Conference Series: Materials Science and Engineering, 928(3), 03-20. Bejaoui, A., Elkhalil, K., Kammoun, A., Alouni, M. S., & Al-Naffouri, T. (2020). Improved design of quadratic discriminant analysis classifier in unbalanced settings. ArXiv Preprint ArXiv:2006.06-355. Ganesalingam, S., Nanthakumar, A., & Ganesh, S. (2013). On Copula based discriminant rule. Journal of Statistics and Management Systems, 16(6), 401-417. Ijomah, M. A. & Biu, O. E. (2017). Variation in the CD4 Counts of HIV/AIDS Patients: A Discriminant Analysis Approach. International Journal of Science and Research, 6(6), 215-221. Li, Y., Liu, B., Yu, Y., Li, H., Sun, J., & Cui, J. (2021). 3E-LDA: Three Enhancements to Linear Discriminant Analysis. ACM Transactions on Knowledge Discovery from Data , 15(4), Nikita, E., & Nikitas, P. (2020). Sex estimation: A comparison of techniques based on binary logistic, probit and cumulative probit regression, linear and quadratic discriminant analysis, neural networks, and naïve Bayes classification using ordinal variables. International Journal of Legal Medicine, 134(3), 1213–1225. Nuhu S. & Gadde S. (2023). Application of Discriminant Analysis to Predict Students’ Performances in Mathematics in Advanced Secondary Schools, European Journal of Statistics, 3(8). Doi: 10.28924/ada/stat.3.8

A New Discrimination Method for Classification of Multivariate Variables: Comparative Analysis

Abstract

Keywords

References

More Articles from INTERNATIONAL JOURNAL OF APPLIED SCIENCES AND MATHEMATICAL THEORY