Income Prediction and Model Explainability using Random Forest, DALEX, and LIME

Project Overview

This project aims to predict the income of individuals based on various demographic and employment-related features using machine learning techniques. A Random Forest model is used for classification, and two powerful explainability tools, DALEX and LIME, are employed to interpret and visualize the model's predictions. This helps in understanding which features are the most influential in determining the income class (e.g., ">50K" or "<=50K").

Dataset

The dataset used in this project is the Adult Income Dataset (also known as the Census Income Dataset), which contains information about individuals and their income. The dataset includes the following columns:

• age: The age of the individual.
• workclass: The type of employment.
• education: The highest level of education attained.
• marital.status: Marital status of the individual.
• occupation: The occupation of the individual.
• relationship: Relationship status.
• race: The race of the individual.
• sex: The gender of the individual.
• native.country: The country of origin.
• income: The target variable, indicating if the individual earns more than 50K a year or not.

The dataset is pre-processed by handling missing values and converting categorical variables to factors.

Libraries Used

• tidyverse: For data manipulation and visualization.
• caret: For model training and evaluation.
• DALEX: For model explainability and feature importance.
• lime: For local interpretable model-agnostic explanations.

Steps Involved

1. Data Loading and Preprocessing:

   • The data is loaded from a CSV file, and the first few rows are displayed.
   • Missing values are checked, and categorical variables are converted into factors.

2. Data Splitting:

   • The dataset is split into training (70%) and testing (30%) sets.

3. Model Training:

   • A Random Forest model is trained using the caret package with 10-fold cross-validation to predict income.

4. Model Evaluation:

   • After the model is trained, predictions are made on the test set, and performance metrics such as Confusion Matrix are calculated.

5. Model Explainability with DALEX:

   • The DALEX package is used to explain the model's predictions. Feature importance is visualized to understand which features play a major role in predicting the income.

6. Model Explainability with LIME:

   • The LIME package provides local explanations for individual predictions. The explanation for the first five test instances is visualized, and the importance of features such as gender is also explored.

Key Visualizations

• Variable Importance: This visualization shows the relative importance of each feature in predicting income.
• LIME Explanations: Local explanations for individual predictions show how specific features influence the outcome for individual data points.

How to Run the Code

To run the code, follow these steps:

1. Install the required packages:

   install.packages("tidyverse")
   install.packages("caret")
   install.packages("DALEX")
   install.packages("lime")