A Hybrid Machine Learning Model Approach to H-1B Visa

Abstract

In recent years immigration has seen a rise, this rise has increased the need for non-immigrant visas for foreign labor workers. One of the most popular in this category is the H-1B visa which has a pretty high rejection rate. Now since the process of H-1B visa is a lottery system, this paper makes an attempt to predict the outcome of this H-1B visa by making use of machine learning models and creating a fusion model for enhancing the results. The machine learning models used in the research are Logistic Regression, Bagging Classifier, SGD Classifier, Gaussian NB, Random Forest, XGB Classifier, AdaBoost Classifier, Gradient Boost Classifier. This paper also emphasizes on finding a pattern between different features and the status of the case. The metrics used for performance analysis are F1 Score, AUC, and Accuracy. The model proposed in this research achieved accuracy, F1-Score, and AUC of 90.79%, 90.58%, and 90.79% respectively