Indian Journal of Science and Technology
Year: 2016, Volume: 9, Issue: 32, Pages: 1-9
Jasjit S. Sodhi1*, Maitreyee Dutta2 and Naveen Aggarwal3
1 Department of Biotechnology, Himachal Pradesh University, Shimla - 171005, Himachal Pradesh, India; [email protected]
2 Department of Computer Science, N.I.T.T.T.R, Chandigarh - 160019, Punjab, India; [email protected]
3 Department of Computer Science, U.I.E.T, Punjab University, Chandigarh - 160014, Punjab, India; [email protected]
*Author for correspondence
Jasjit S. Sodhi
Department of Biotechnology
Objectives: This paper presents the use of machine learning technique in order to eliminate the components affecting the human decision making process. To assist decision making in career counseling, an Artificial Intelligence model was implemented using Artificial Neural Network (ANN) in MATLAB for predicting vocational stream of pursuit based on the behavioral characteristic of the beneficiary. Methods/Statistical Analysis: The Differential Aptitude Test (DAT) battery; and Scientific Knowledge and Aptitude Test (SKAT) were used to assess an individual’s specific abilities in different areas. The training data set for the ANN model was procured in the form of normalized scores based on occupational/vocational profiles as given by the authors of DAT battery and SKAT. The trained ANN was tested with normalized stanine scores of 100 tenth class studentsraised through cluster random sampling technique for predicting a vocational stream of pursuit. In order to evaluate its accuracy and efficiency, three techniques of classification were employed. The unclassified data was classified by using Discriminant function analysis, ANN and two classifications were obtained from the trained counselors. Findings: The classification result obtained from the above mentioned techniques were compared and was found that the ANN system and Discriminant function analysis agreed approximately by 91% over all the test cases. The results of the statistical method support the classification made by the ANN. The two counsellors were in agreement with ANN’s classification output by approximately 81%. However, the counsellors disagreed with each other’s prediction approximately by 27% over all the test cases. The experimental results support the hypothesis that the proposed machine learning technique performed better than the prediction made by counsellors. Application/Improvements: In the Indian scenario, the developed machine learning system may be used as a standalone system in places where there is the paucity of counsellorsor can assist in the career decision making provided by human beings.
Keywords: Artificial Neural Network, Career Counseling, Decision Support System, Differential Aptitude Test Battery, Discriminant Function Analysis, Scientific Aptitude and Knowledge Test
Subscribe now for latest articles and news.