Predicting Adult Census Income Using XGBoost Gradient Boosted Trees System

Haibin Lai 赖海斌 12211612

In this project, We try to predict Adult Census Income Dataset by using XGBoost Gradient Boosted Trees System.

This directory contains 3 main programs:

• AutogluonProcess.ipynb & AutogluonProcess.py: These files are used to run Autogluon and help us decide which model to use.

• XGBoost.ipynb: This file is used to run XGBoost.

• Visualization.ipynb: This file is for data visualization.

Running Preparation

Needed Python Version:

! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! !

python <=3.11 (package autogluon can't support python 3.12!)

! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! !

My computer runs well in 3.11

Needed Package:

The following version work well in Anaconda:

Library	Version	Notes
autogluon	1.1.0	自动化机器学习包
pandas	2.2.1	处理数据
warnings	in Python	-
matplotlib	3.8.4	画图
seaborn	0.12.0	数据可视化
sklearn	1.4.0	决策树框架 (1.5.0 tested OK)
xgboost	2.0.3	提供 XGBoost 分类器
jupyter	1.0.0

Using xgboost:

XGBoost: eXtreme Gradient Boosting library. Contributors: https://github.com/dmlc/xgboost/blob/master/CONTRIBUTORS.md

Installing

Pip 21.3+ is required

pip install xgboost

Installing autogluon

(tips: install time is a bit long! And we don't fully need it!)

这个包有很多版本限制，它可能会改掉一些下载的东西（比如它会卸载scikit-learn 1.4.2，然后重新安装个版本）要依赖很多包（比如torch），文件非常大，并且需要管理员权限来安装，我们只在Autogluon文件中进行了跑， 并且最终提交没有依赖那两个文件，而是XGBoost文件，如果跑不起来也没问题。 另外很神奇的是，这个包的预测跟它的版本有关

pip install autogluon

Useful website

This project's Github website: https://github.com/Laihb1106205841/CS311-AI-Peoject3-XGBoost.git

XGBoost Introduction: https://xgboost.readthedocs.io/en/stable/tutorials/model.html

XGBoost Github website: https://github.com/dmlc/xgboost?tab=security-ov-file