Smart Career Guidance System

Abstract

We have developed a career guidance system that helps those students who are about to begin their higher education. Most of the time, students are not aware of what career path to follow or which academic major is in accordance with their interests. The system analyzes students' skills, abilities, and interests and recommends the five fields which are most suitable for them. This project helps students identify a specific domain that fits their skills and interests. Smart Career Guidance System is a web-based application built on the Django framework. We have deployed various Machine Learning techniques and algorithms to mimic a one-on-one meeting with an experienced career counselor. The data was collected in the form of a questionnaire that was based on Holland Occupational Themes and the Theory of Multiple Intelligences. A total of 392 graduates completed this online survey. SMOTE oversampling is used to evaluate the machine learning classifiers since the data is highly imbalanced. We tested XGBoost and Random Forest classifiers for recommending the best-suited career options which furnish AUC-ROC performance scores of 0.9952 and 0.9963 respectively. A fine-tuned version of the Random Forest Classifier has successfully attained an AUC-ROC performance score of 0.9976 which indicates the minimal false-positive rate. Ms. Areej Kamal, Ms. Hadia Rafiq and Ms. Batul Naushad have collaboratively conducted all activities of the project including data collection and cleaning, literature review, testing of ML models and development of the final solution. Mr. Shahab Tahzeeb directed and supervised all phases of the project with his immense knowledge and expertise.