All projects

Machine learning

Predictive modeling.

I trained regression and classification models, then checked where their predictions went wrong.

  • Python
  • scikit-learn
  • pandas
  • NumPy
Try the example

Interactive example · Sample data

Move the cutoff.

These 70 synthetic samples have positive or negative labels. Move the decision threshold to see how precision and recall change.

Illustration of the project’s sample exercise
Enable JavaScript to try this example.

How I built it.

Coursework

Machine Learning, Data Analytics & Advanced Programming for Data Analysis

Prepare the features

I cleaned the data, handled missing values, encoded categorical variables, and scaled features to prepare them for training.

Train the models

I used scikit-learn to fit supervised models for both numerical prediction and classification tasks.

Inspect the errors

For classification, I used confusion matrices, precision, and recall alongside accuracy. I revisited feature selection as I compared the results.