Measuring and comparing the efficiency of logistic regression, decision tree and SVM with k-means algorithm in college recommendation system

• Published in: AIP conference proceedings Vol. 2853; no. 1
• Main Authors: Kumar, M. Vinod, Lakshmi, S. Vidhya, Sabitha, R.
• Format: Journal Article Conference Proceeding
• Language: English
• Published: Melville American Institute of Physics 07.05.2024
• Subjects: Algorithms; Classification; Classifiers; Colleges & universities; Datasets; Decision trees; Performance evaluation; Recommender systems; Regression; Regression analysis; Support vector machines
• ISSN: 0094-243X; 1551-7616
• DOI: 10.1063/5.0198460

The purpose of this research is to evaluate the performance of four distinct approaches to prediction—namely, Logistic Regression (LR), Decision Tree (DT), K Means Algorithm (KMA), and Support Vector Machine (SVM) Classification—in order to determine which one yields the most accurate results. The Following are the Components and Methods: When testing the Classification technique, a dataset consisting of 778 items is employed. In the subject of education, the Logistic Regression (LR), K-Means, Decision Tree, and Support Vector Machine (SVM) are all viable frameworks that have been suggested and developed for the College Recommendation system. K-Means was the first of these frameworks to be built. Through the use of G power, we were able to establish that we need 55 people for each of the conditions. Both the precision and accuracy of the classifiers have been analysed and recognised. Through the use of G Power, we were able to establish that each of our groups required 55 individuals. The College dataset had a sample size of 303 students, 76 attributes, and a few blanks in addition to those numbers. A clinical study was conducted to establish the appropriate size of the sample, and the following were the findings: 80 percent predictive power on the pretest, 0.05 alpha, and a 1. When applied to the College Recommendation System data set, the results demonstrate that the K-Means classifier produces the same group that it predicts (50 percent). This is also the case for Logistic regression (86 percent), Decision Tree (88 percent), K-Means (50 percent), and Support Vector Machine (50 percent) (85 percent). The value assigned to significance is zero. Consequently, DT performs better than LR, SVM, and K-Means. The findings indicate that Decision Tree offers greater performance in terms of precision and accuracy when compared to SVM, LR, DT, and K-Means. This is shown by the fact that Decision Tree.