Recent years have witnessed a surge in Computer Science (CS) research, which achieved unparalleled successes in a kaleidoscope of science and engineering applications spanning artificial intelligence [1], natural language processing [2], computer vision [3], statistical learning theory [4], blockchain and cryptocurrency [5], computational biology [6] and bioinformatics [7], intelligent grids and Internetof-Things (IoT) [8]. Consequently, the difficulty of pursuing a higher degree in CS or related subjects in decent graduate schools has been elevated to an unprecedented new height [9]. In this paper, motivated by these observations, the authors, consisting of a data-driven earth scientist, a computational biologist, and a machine learning researcher, study the factors governing the admissions of graduate schools in the U.S. by means of Generalized Linear Models (GLM), Generalized Additive Models (GAM), distribution-free methods such as Ensemble Learning (EL), and Discrete Bayesian Networks (DBN). We seek answers to several interesting questions and render crisp insights from the model inference and analysis. We hope this paper may inspire future graduate schools applicants. We finally conclude this paper by identifying limitations that are not addressed by this paper and pointing out several possible future directions.
We use the paper draft philosophy introduced by Li-Yi Wei throughout this project. The report can be found online here.
Chen Liu, Zehao Su, and Jiayao Zhang. All authors contributed equally.
If you found this repo useful, please consider cite this project.
A possible bibtex entry may be:
@misc{LSZ18,
author = {Liu, Chen and Su, Zehao and Zhang, Jiayao}
title = {Understanding Admission Results of CS Graduate Programs in U.S. Universities},
year = {2018},
publisher = {GitHub},
journal = {GitHub repository},
howpublished = {\url{https://github.com/zjiayao/stat7614}},
}