IIITD, CSE343 Machine Learning Monsoon 2020 Project

With another fresh change in the Immigration policies adopted by USA during the pandemic, a focus has been bought on the state on the H1B visa. In 2019, over 150,000 people applied fresh, and near a million applicants applied for renewal. The immigration process is very closed process, as no details are provided to applicants on which basis the selection/rejection has been made. However, with this dataset release, which has been anonymized, we can now taking a deep dive into what their policies revolve around.

We want to analyse and see which factors has the US been caring about, and which traits downplay a profile. We also want to learn how well machine learning methods can predict the selection procedure.

Using the analysis, one can potentially reduce a lot of risk involved in the application process, both to independent applicants and employers who have funded the visa application.

Data set description : Using official statistics we were able to find the datasets for H1B approval , and we selected the dataset for the year 2019 . There were over 260 features recorded in that, and the target variable was CASE_STATUS. Some of the major attributes included Job title , employer name , wage , employment duration, renewal and so on…

A complete report of all findings and model performances can be viewed here