In this project, We have performed a detailed comparison of 12 different classification models for the Adult Census Income dataset to predict whether a person's income is over 50000 USD or not. We have used 12 different classification models and performed different preprocessing techniques on the dataset and hyperparameter tuning for the models to further improve the accuracy in the prediction of labels mentioned.
We have observed 86% overall accuracy for default values and improvement of over 2% (88.3% overall accuracy) after the optimization of the models. The results of the experiments show that a similar set of preprocessing and classification models can be used for comparing many prediction problems similar to the one presented in the project.
The repository contains 4 files :
- Project_Report - Report file containing all my findings
- Project_Code.py - Project code
- data.csv - Dataset used for the project
- Experiments - Text file containing the experiments performed within the project