Costa Rican Household Poverty Level Prediction

Analysis based on a dataset of Costa Rican household characteristics in order to improve the Proxy Means Test (or PMT) performance.

Here's the Kaggle page for this challenge.

---

My Approach:

The Kernel is divided into 3 parts:

Data Exploration:

Here I try to analyse the Target feature. I visualise the relationship of the Target variable with other features and use the graphs to draw conclusions.

I also identified the types of features and cleaned up the inconsistency in data here.

Preprocessing:

I do 4 important things over here -

• missing values imputation
• generating ordinal features from dummy features
• removing redundant features
• creating new household-wide features

Modelling:

I try 2 gradient boosting machines - LightGBM and XGBoost - to model the data.

I have used a 10-fold cross-validation strategy to get the CV scores of each model for comparision.

In the end, I decided on the LightGBM model and got an F1-macro score of 0.431 on the test data (currently in the top 10% of Kaggle leaderboard).

---

I have included the data in the ./input/ directory and my kernel is present in the ./notebook/ directory