FrankDTS/Predict-People-Salary

Built several Machine Learning (Logistic Regression, Random Forest, and XGboost) to predict whether a person makes over 50K a year

Predict Salary Project OverView

• Built several Machine Learning models (Logistic Regression, RandomFordst, and XGboost) to predict whether a person makes over 50K a year
• Several Feature Engineering methods to fill with columns that have NA values.

Code and Resourced Used

• Python 3.8
• Packages: pandas, numpy, seaborn, matplotlib, sklearn, Xgboost
• [Xgboost parameter] (https://xgboost.readthedocs.io/en/latest/parameter.html)
• [Data] (https://archive.ics.uci.edu/ml/index.php)

Data Preprocessing

1. Normalize or use logistic to transform the numeric columns (age, fnlwgt, education_number, hours)

2. Group approximate equal columns to groups to reduce the dimension (workclass, martial_status, native_country)

   1. workclass: replace 'Without pay' and 'Never-worked' classes to 'Non-pay' class
   2. martial_status: replace 'Divorced' class to 'Seperated' class
   3. native_country: use continent to group each country.

3. one-hot encoding for categorical parameters

4. Built Xgboost model to predict workclass, occupation, and native_country's Null value

Fit and Predict models

1. Logistic model
2. RandomForest model
3. Xgboost model