Binary Classification for Teacher Donor’s Project

DOI

10.2991/iceess-18.2018.40How to use a DOI?

Keywords

binary classification, natural language processing, statistical machine learning models, Python

Abstract

Classification always plays an important role in statistical machine learning, which contains both binary classification problems and multi-label classification problems. This article focuses on binary classification models including natural language processing for text objects to help teachers to improve their chances of being funded based on real data sets collected by DonorsChoose.org. Comparing about two natural language processing methods for projects proposals proposed by teachers, we also implement various statistical algorithms on our data sets, aiming to enhance the classification accuracy which can be measured by model accuracy and the area under the curve(AUC). In conclusion, the text objects are important for computer to conduct supervised learning and the length of the proposal and the price column are the crucial features. In addition, the best model will be the LightBGM with AUC 0.77 and accuracy 86%.

Copyright

© 2018, the Authors. Published by Atlantis Press.

Open Access

This is an open access article distributed under the CC BY-NC license (http://creativecommons.org/licenses/by-nc/4.0/).

Download article (PDF)

TY  - CONF
AU  - Yunwei Zhang
AU  - Zibin Zhang
PY  - 2018/10
DA  - 2018/10
TI  - Binary Classification for Teacher Donor’s Project
BT  - Proceedings of the 2018 International Conference on Education, Economics and Social Science (ICEESS 2018)
PB  - Atlantis Press
SP  - 157
EP  - 160
SN  - 2352-5398
UR  - https://doi.org/10.2991/iceess-18.2018.40
DO  - 10.2991/iceess-18.2018.40
ID  - Zhang2018/10
ER  -