Convolutional Neural Network Based Career Recommender System for Pakistani Engineering Students
Takreem Saeed, Muhammad Sufian, Mubashir Ali, Attique Ur Rehman · 2021 International Conference on Innovative Computing (ICIC) · 2021
In recent years, Recommender systems are utilized in a variety of areas. One reason behind why we want a recommender system in current society is that an individual has many alternatives to use because of the pervasiveness of the Internet. A recommender system seeks to estimate and predict user content preference. Old recommender systems used State-of-the-art recommender algorithms like content based filtering to predict ratings. Career Recommender system provides Engineering candidates the best possible available jobs relevant to their skills, qualification, etc. Four to six major engineering disciplines are covered in this recommender system. The proposed approach is tested using a career recommendation dataset which is collected from many students of different disciplines of various universities. A deep NLP based CNN model is used to predict the best jobs with maximum precision. 512 hidden layers are used to increase the performance of this system. Career recommendation takes care of the users and saves their cost and time spending on traditional job searching methods. Comparative study demonstrates that the proposed methodology of prediction of the best jobs achieves better results with an accuracy of 84% when matched with content based filtering technique where 81% accuracy is gained for content based career recommender system.