Create test/train splits and build your ALS model

You already know how to build an ALS model, having done it in the previous chapter. We will do that again, but we'll take some additional steps to fully build out a cross-validated model.

First, let's import the requisite functions and create our train and test data sets in preparation for the cross validation step.

This exercise is part of the course

Building Recommendation Engines with PySpark

Exercise instructions

Import the RegressionEvaluator from ml.evaluation, the ALS algorithm from ml.recommendation, and the ParamGridBuilder and the CrossValidator from ml.tuning.
Create an .80/.20 train/test split on the ratings data using the randomSplit method. Name the datasets train and test, and set the random seed to 1234.
Build out the ALS model, telling Spark the names of the columns in the ratings dataframe that correspond to the userCol, itemCol and ratingCol. Set the nonnegative argument to True, the coldStartStrategy to "drop" and let Spark know that these are not implicitPrefs by setting the implicitPrefs argument to False. Call this model als.
Verify that the model was created by calling the type() function on als. The output should indicate what type of model it is.

Hands-on interactive exercise

Have a go at this exercise by completing this sample code.

# Import the required functions
from pyspark.ml.evaluation import ____
from pyspark.ml.recommendation import ____
from pyspark.ml.tuning import ____, ____

# Create test and train set
(train, test) = ratings.___([0.____, 0.____], seed = ____)

# Create ALS model
als = ALS(userCol="____", itemCol="____", ratingCol="____", nonnegative = ____, implicitPrefs = ____)

# Confirm that a model called "als" was created
type(____)

Edit and Run Code

This exercise is part of the course

Building Recommendation Engines with PySpark

AdvancedSkill Level

4.8+

Start Course for Free

This chapter will show you how powerful recommendations engines can be, and provide important distinctions between collaborative-filtering engines and content-based engines as well as the different types of implicit and explicit data that recommendation engines can use. You will also learn a very powerful way to uncover hidden features (latent features) that you may not even know exist in customer datasets.

Exercise 1: Why learn how to build recommendation engines?Exercise 2: See the power of a recommendation engine Exercise 3: Power of recommendation engines Exercise 4: Recommendation engine types and data types Exercise 5: Collaborative vs content-based filtering Exercise 6: Collaborative vs content based filtering part II Exercise 7: Implicit vs explicit data Exercise 8: Ratings data types Exercise 9: Uses for recommendation engines Exercise 10: Alternate uses of recommendation engines.Exercise 11: Confirm understanding of latent features

In this chapter you will review basic concepts of matrix multiplication and matrix factorization, and dive into how the Alternating Least Squares algorithm works and what arguments and hyperparameters it uses to return the best recommendations possible. You will also learn important techniques for properly preparing your data for ALS in Spark.

Exercise 1: Overview of matrix multiplication Exercise 2: Matrix multiplication Exercise 3: Matrix multiplication part II Exercise 4: Overview of matrix factorization Exercise 5: Matrix factorization Exercise 6: Non-negative matrix factorization Exercise 7: How ALS alternates to generate predictions Exercise 8: Estimating recommendations Exercise 9: RMSE as ALS alternates Exercise 10: Data preparation for Spark ALS Exercise 11: Correct format and distinct users Exercise 12: Assigning integer id's to movies Exercise 13: ALS parameters and hyperparameters Exercise 14: Build out an ALS model Exercise 15: Build RMSE evaluator Exercise 16: Get RMSE

In this chapter you will be introduced to the MovieLens dataset. You will walk through how to assess it's use for ALS, build out a full cross-validated ALS model on it, and learn how to evaluate it's performance. This will be the foundation for all subsequent ALS models you build using Pyspark.

Exercise 1: Introduction to the MovieLens dataset Exercise 2: Viewing the MovieLens Data Exercise 3: Calculate sparsity Exercise 4: The GroupBy and Filter methods Exercise 5: MovieLens Summary Statistics Exercise 6: View Schema Exercise 7: ALS model buildout on MovieLens Data Exercise 8: Create test/train splits and build your ALS model

Current Exercise

Exercise 9: Tell Spark how to tune your ALS model Exercise 10: Build your cross validation pipeline Exercise 11: Best Model and Best Model Parameters Exercise 12: Model Performance Evaluation Exercise 13: Generate predictions and calculate RMSE Exercise 14: Interpreting the RMSE Exercise 15: Do recommendations make sense

In most real-life situations, you won't not have "perfect" customer data available to build an ALS model. This chapter will teach you how to use your customer behavior data to "infer" customer ratings and use those inferred ratings to build an ALS recommendation engine. Using the Million Songs Dataset as well as another version of the MovieLens dataset, this chapter will show you how to use the data available to you to build a recommendation engine using ALS and evaluate it's performance.

Exercise 1: Introduction to the Million Songs Dataset Exercise 2: Confirm understanding of implicit rating concepts Exercise 3: MSD summary statistics Exercise 4: Grouped summary statistics Exercise 5: Add zeros Exercise 6: Evaluating implicit ratings models Exercise 7: Specify ALS hyperparameters Exercise 8: Build implicit models Exercise 9: Running a cross-validated implicit ALS model Exercise 10: Extracting parameters Exercise 11: Overview of binary, implicit ratings Exercise 12: Binary model performance Exercise 13: Recommendations from binary data Exercise 14: Course recap