Classification of employable students under principal components analysis

• Published in: AIP conference proceedings Vol. 2724; no. 1
• Main Author: Akella, Suresh
• Format: Journal Article Conference Proceeding
• Language: English
• Published: Melville American Institute of Physics 28.04.2023
• Subjects: Big Data; Chemical reactions; Eigenvalues; Eigenvectors; Engineering; Engineering education; Parameters; Principal components analysis; Software; Students; Vector analysis
• ISSN: 0094-243X; 1551-7616
• DOI: 10.1063/5.0130303

Big data with multiple parameters that affect the variation in outcome happens in stock market, mechanical processes, weather, chemical processes etc. The Principal Components Analysis, PCA, an unsupervised analysis is a simple statistical and numerical tool to reduce the multidimensional space of occurrence of the data in say m parameters to the first 2 or 3 principal axes which contribute to more than 70% of the outcome variations. In this study a group of 154 employed students who completed their 3rd year are evaluated first reducing the variables from 25 subjects they studied to five subject's one critical subject of first five semesters. Their internal marks were taken for Eigen value and Eigen vector analysis. The first two principal components were identified, and the transformed data is seen in this representative 2D space. A small lower order cluster is observed which can be improved in future Pedagogy and Campus Recruitment Training, CRT, for employment of the students to better companies. Though all subjects in engineering have equal importance the performances in the identified subjects reflect their total ability. As the application is on Engineering students, internal marks matrix is termed as Knowledge matrix, eigen values as Engineering Values and eigen vector as Engineering Vector to showcase the application. MS-Excel and MATLAB software tools were used for numerical calculations.