Statistics for Machine Learning

Hyperparameter selection in recommendation engines using grid search

In any machine learning algorithm, the selection of hyperparameter plays a critical role in how well the model generalizes the underlying data. In a similar way, in recommendation engines, we have the following hyperparameter to play with:

• Number of iterations: The higher the number of iterations, the better the algorithm converges. In practice, it has been proven that ALS converges within 10 iterations, but readers are recommended to try various values and see how the algorithm works.
• Number of latent factors: These are the explanatory variables which try to provide reasons behind crowd behavior patterns. The higher the latent factors, the better the model, but too high ...