A Machine Learning and Deep Learning Approach for Selecting the Most Suitable Teacher at the Ministry of Education رسالة ماجستير

Abstract

There is a significant change that facing the recruitment of teachers for the Ministry of Education in Palestine, there is a limited number of vacancies and a higher number of applicants, however, the implants between the availability of opportunities and the supply of qualified candidates may create hard selection how is the suitable candidates, and leads to inefficiencies in the hiring process so addressing the issue is cru\tical for ensuring that the right person in the right, place. Moreover, the Palestinian Ministry of Education offers a few positions for teachers every year, and many of the candidates who try to fill the empty positions in the research applications were more than 50000 applications for just almost 1000 positions Our study offers using ML to facilitate the selection criteria and choose the most appropriate candidate[1]. In this study we use ML models and compare imbalance and balanced datasets by training the models on the candidate variables such as academic performance, teaching experience gender, demographic factors, and more, as well as classification of the participants into three categories absent, not place, and place, we processed the datasets using support vector machines (SVM), decision tree (DT), XGBoost, Random forest (RF), gradient Boost (GB), Adaboost, Naive Bayes, logistic regression, KNN boost, and Multilayer Perceptron Neural Networks (MLPNNs)[2]The performance of each model was evaluated, and the random forced model was selected because it achieved the highest accuracy, 99%, for both imbalance and balance datasets.