Before building our ALS model, we need to split the data into training data and test data. Use the randomSplit() method to split the ratings dataframe into training_data and test_data using an 0.8/0.2 split respectively and a seed for the random number generator of 42.
Tell Spark which columns contain the userCol, itemCol and ratingCol. Use the .columns method if needed. Complete the hyperparameters. Set the rank to 10, the maxIter to 15, the regParam or lambda to .1, the coldStartStrategy to "drop", the nonnegative argument should be set to True, and since our data contains explicit ratings, set the implicitPrefs argument to False.
Now fit the als model to the training_data portion of the ratings data by calling the als.fit() method on the training_data provided. Call the fitted model model.
Generate predictions on the test_data portion of the ratings data by calling the model.transform() method on the test_data provided. Call the predictions test_predictions. Feel free to view the predictions by calling the .show() method on the test_predictions

Estimating recommendations