FUZZY-FILTER FEATURE SELECTION FOR ENVISIONING THE EARNINGS OF HIGHER EDUCATION GRADUATES

Published on Jan 1, 2018
Tahseen A. Wotaifi , Eman S. Al-Shamery1
Estimated H-index: 1
Source
Abstract
With the great change in the labor market and contemporary education, a high consideration is paid for the issue of graduates’ employability and their earning. However, educational institutes and universities have different polices for preparing graduates to compete strongly in the labor market based on their academic skills and experience. This paper aims at (1) predicting the earnings of alumni for six years after graduation, and (2) identifying the most important factors that directly affect the earnings. Hence, the accuracy of the prediction can be improved based on the higher education system in the United States. Unlike previous research, this study contributes to the use of fuzzy logic on three filter methods: Relief Attribute Evaluation, Correlation Attribute Evaluation, and classifier Attribute Evaluation. Because these three methods provide different weight to the same attribute, the fuzzy logic is used to obtain a single weight. The proposed system is applied on higher education of United States; specifically the college scorecard data set that contains nearly (8000) educational institutions and (1825) feature for each. The proposed Fuzzy-Filter method has selected 45 features only. Accordingly, two models are used to predict graduates' earnings, which are Random Forest and Linear Regression with Mean Absolute Error (MAE) (0.055) and (0.065) respectively. The research findings are better in terms of Mean Absolute Error (MAE) and reducing the number of features in comparison to previous studies.
References0
Newest
Cited By1
Newest
The competition between scientific institutions is increased every day. Every institution tends to improve its reputation by producing and publishing high-quality scientific research. Clustering and evaluating the educational institutions are important for professors, policymakers, as well as students. This research aims to develop a Jarvis-Patrick algorithm for scientific institutions clustering, which is one of the graph-based techniques. It suffers from the problem of a large number of cluste...
Source