Adult-Income

Exploring the possibility in predicting income level based on the individual’s personal information

William Rodemoyer

Data

https://www.kaggle.com/datasets/wenruliu/adult-income-dataset

An individual’s annual income results from various factors. Intuitively, it is influenced by the individual’s education level, age, gender, occupation, and etc.

Data Analysis Visuals

Of the individuals who make greater than 50k, clearly are older in age on average.

!

Both Genders who make more than 50k, work more hours per week than those who make 50k or less.

On average, males work more hours per week compared to females when comparing their respectable income categories.

Models

I used 2 types of models with multiple variants to predict the income

• Logistic Regression Model
  • Base
  • Tuned
  • Over Sample
  • Under Sample
  • PCA
• KNN Model

  • Base

  • Tuned

  • Over Sample

  • Under Sample

  • PCA

     Model Name       Precision    Recall      F1 Score     Accuracy
    
     LR Tuned Train   0.737003     0.602072    0.662740     0.852826
     
     LR Tuned Test    0.732665     0.605663    0.663138     0.853815
    
     
     Model Name       Precision    Recall      F1 Score     Accuracy
                                        
     KNN Tuned Train   0.773205    0.586132    0.666796     0.859307
    
     KNN Tuned Test    0.725936    0.562500    0.633852     0.845611
    

Final Model Results

After comparing the models, my recommendation is:

• Logistic Regression Model w/ OverSampling
• This model gave us an 81% chance of correctly predicting an individuals income.
• We did have other models that scored higher than 81%
• But this model was the most balanced.

For Further Informtion

For any additional questions, please contact wrodemoyer@gmail.com