A comparative Study on Graduates' Employment in Malaysia by using Data Mining

• Published in: Journal of Physics: Conference Series
• Main Author: Binti A'Rifian N.I.N.; Binti Mohd Daud N.S.A.; Muhamad Romzi A.F.B.; Binti Md Shahri N.H.N.
• Format: Conference paper
• Language: English
• Published: Institute of Physics Publishing 2019
• Online Access: https://www.scopus.com/inward/record.uri?eid=2-s2.0-85076111275&doi=10.1088%2f1742-6596%2f1366%2f1%2f012120&partnerID=40&md5=fbe38dbc1e0fbc70dbe5b2a18856de88

This study implements data mining to extract knowledge by analysing the graduates' employment dataset from year 2017 obtained from Ministry of Higher Education (MoHE). The objective of this study is to compare three predictive models which are Decision Tree (DT), Logistic Regression (LR) and Artificial Neural Network (ANN). Besides, this study is also done to determine the best predictive model for predicting graduates' employment sectors whether in public sector or private sectors. Every graduate student wishes to choose the right path in determining which sectors they are going to be entered, either to the public or private sectors. Usually, most graduates in Malaysia prefer the employment in the public sector rather than the private sector. Using data mining to discover the relationship and patterns can help in making a better decision. Prediction model is a must to determine the best performance when dealing with the large data set which helps the graduates to choose a sector based on the type of data or information that he/she furnishes. Based on the analysis, Artificial Neural Network (ANN 5) is the best model in predicting placement of employed graduates whether in public sector or private sector compared to the other models. ANN 5 is the highest accuracy at 81.52% and sensitivity at 65.67% while for the specificity of ANN 5 is 91.44%. The misclassification rate of ANN 5 is 18.48% which is the lowest compared to the other models. Overall, ANN 5 is the best model to predict negative target which is graduates employed in private sector since the value of specificity is higher than sensitivity. The result of this study can be used by government, universities and other responsible agencies in order to predict whether graduates will be employed in public or private sectors. © Published under licence by IOP Publishing Ltd.