Nowadays, with the increase in data production speed, the process of data analysis has faced many problems because this big data is often accompanied by plug-in data and redundant data. Therefore, the use of dimensional methods in the pre-data analysis stage is necessary. In data mining, dimensional reduction is one of the most important steps in data pre-processing. Principal component analysis (PCA) and linear discriminant analysis (LDA) are often used to reduce dimensions in data mining. The LDA method is a monitored and controlled method but the PCA is not controlled method. When the number of samples in classes is large and when training data is uniformly distributed, LDA works better. LDA is a traditional statistical method for classification. In this research, we improve the LDA method by identifying and removing redundant variables. Then we use this improved LDA method to classify the diagnosis data of Covid-19 patients.
Vajargah, Kianoush Fathi; Golshan, Hamid Mottaghi; and Farahabadi, Fazel Badakhshan
(R2025) Improving the LDA Linear Discriminant Analysis Method by Eliminating Redundant Variables for the Diagnosis of COVID-19 Patients,
Applications and Applied Mathematics: An International Journal (AAM), Vol. 18,
1, Article 8.
Available at: https://digitalcommons.pvamu.edu/aam/vol18/iss1/8