Regression and Logistics Model

With practically two lines, we built a Logistic Regression model with 80% accuracy of the training dataset.

The .fit() method optimizes the objective function of the problem and stores the “ideal” weights inside the variable itself, becoming an attribute.

We can predict the test dataset with the .predict() method.

I took the opportunity and submitted the csv file to Kaggle to compare the performance:

The model managed to achieve a score of 75%. From what I’ve been studying, it’s not uncommon for us to underperform when submitting predictions, as our model was trained 100% on top of the training set, which means it has a better fit to the dataset train.