Lab 3: Matrix Factorisation for Collaborative Filtering in Recommender Systems

COM6012 Scalable Machine Learning 2021 by Haiping Lu at The University of Sheffield

Study schedule

Task 1: To finish by Wednesday. Essential
Task 2: To finish by Thursday. Exercise
Task 3: To explore further. Optional

PySpark APIs in pictures (highly recommended)

Learn PySpark APIs via Pictures (from recommended repositories in GitHub, i.e., found via recommender systems!)

1. Movie recommendation via collaborative filtering

Get started

Let's start a pyspark shell with 2 cores using a regular queue:

qrshx -pe smp 2 # request 2 CPU cores without using our reserved queue
source myspark.sh # myspark.sh should be under the root directory
cd com6012/ScalableML # our main working directory
pyspark --master local[2] # start pyspark with 2 cores requested above.

Collaborative filtering

Collaborative filtering is a classic approach for recommender systems. These techniques aim to fill in the missing entries of a user-item association matrix primarily based on the matrix itself. spark.ml currently supports model-based collaborative filtering, in which users and products are described by a small set of latent factors that can be used to predict missing entries, using the alternating least squares (ALS) algorithm.

API: class pyspark.ml.recommendation.ALS(rank=10, maxIter=10, regParam=0.1, numUserBlocks=10, numItemBlocks=10, implicitPrefs=False, alpha=1.0, userCol='user', itemCol='item', seed=None, ratingCol='rating', nonnegative=False, checkpointInterval=10, intermediateStorageLevel='MEMORY_AND_DISK', finalStorageLevel='MEMORY_AND_DISK', coldStartStrategy='nan', blockSize=4096)

The following parameters are available:

rank: the number of latent factors in the model (defaults to 10).
maxIter is the maximum number of iterations to run (defaults to 10).
regParam: the regularization parameter in ALS (defaults to 0.1).
numUserBlocks/numItemBlocks: the number of blocks the users and items will be partitioned into in order to parallelize computation (defaults to 10).
implicitPrefs: whether to use the explicit feedback ALS variant or one adapted for implicit feedback data (defaults to false which means using explicit feedback).
alpha: a parameter applicable to the implicit feedback variant of ALS that governs the baseline confidence in preference observations (defaults to 1.0).
nonnegative: whether or not to use nonnegative constraints for least squares (defaults to false).
coldStartStrategy: can be set to “drop” in order to drop any rows in the DataFrame of predictions that contain NaN values (defaults to "nan", assigning NaN to a user and/or item factor is not present in the model.
blockSize: the size of the user/product blocks in the blocked implementation of ALS (to reduce communication).

Movie recommendation

In the cells below, we present a small example of collaborative filtering with the data taken from the MovieLens project. Here, we use the old 100k dataset, which has been downloaded in the Data folder but you are encouraged to view the source.

The dataset looks like this:

    196     242     3       881250949
    186     302     3       891717742
    22      377     1       878887116
    244     51      2       880606923

This is a tab separated list of

    user id | item id | rating | timestamp

Explicit vs. implicit feedback

The data above is typically viewed as a user-item matrix with the ratings as the entries and users and items determine the row and column indices. The ratings are explicit feedback. The Mean Squared Error of rating prediction can be used to evaluate the recommendation model.

The ratings can also be used differently. We can treat them as numbers representing the strength in observations of user actions, i.e., as implicit feedback similar to the number of clicks, or the cumulative duration someone spent viewing a movie. Such numbers are then related to the level of confidence in observed user preferences, rather than explicit ratings given to items. The model then tries to find latent factors that can be used to predict the expected preference of a user for an item.

Cold-start problem

The cold-start problem refers to the cases when some users and/or items in the test dataset were not present during training the model. In Spark, these users and items are either assigned NaN (not a number, default) or dropped (option drop).

MovieLens100k

Let's study ALS for movie recommendation on the MovieLens 100K Dataset.

from pyspark.ml.evaluation import RegressionEvaluator
from pyspark.ml.recommendation import ALS
from pyspark.sql import Row

Read the data in and split words (tab separated):

lines = spark.read.text("Data/MovieLens100k.data").rdd
parts = lines.map(lambda row: row.value.split("\t"))

We need to convert the text (String) into numbers (int or float) and then convert RDD to DataFrame:

ratingsRDD = parts.map(lambda p: Row(userId=int(p[0]), movieId=int(p[1]),rating=float(p[2]), timestamp=int(p[3])))
ratings = spark.createDataFrame(ratingsRDD).cache()

Check data:

ratings.show(5)
# +------+-------+------+---------+
# |userId|movieId|rating|timestamp|
# +------+-------+------+---------+
# |   196|    242|   3.0|881250949|
# |   186|    302|   3.0|891717742|
# |    22|    377|   1.0|878887116|
# |   244|     51|   2.0|880606923|
# |   166|    346|   1.0|886397596|
# +------+-------+------+---------+
# only showing top 5 rows

Check data type:

ratings.printSchema()
# root
#  |-- userId: long (nullable = true)
#  |-- movieId: long (nullable = true)
#  |-- rating: double (nullable = true)
#  |-- timestamp: long (nullable = true)

Prepare the training/test data with seed 6012:

myseed=6012
(training, test) = ratings.randomSplit([0.8, 0.2], myseed)
training = training.cache()
test = test.cache()

Build the recommendation model using ALS on the training data. Note we set cold start strategy to drop to ensure we don't get NaN evaluation metrics:

als = ALS(userCol="userId", itemCol="movieId", seed=myseed, coldStartStrategy="drop")
model = als.fit(training)
# 21/02/19 09:34:30 WARN BLAS: Failed to load implementation from: com.github.fommil.netlib.NativeSystemBLAS
# 21/02/19 09:34:30 WARN BLAS: Failed to load implementation from: com.github.fommil.netlib.NativeRefBLAS
# 21/02/19 09:34:31 WARN LAPACK: Failed to load implementation from: com.github.fommil.netlib.NativeSystemLAPACK
# 21/02/19 09:34:31 WARN LAPACK: Failed to load implementation from: com.github.fommil.netlib.NativeRefLAPACK

The warnings on BLAS and LAPACK are about optimised numerical processing. The warning messages mean that a pure JVM implementation will be used instead of the optimised ones, which need to be installed separately. We are not installing them in this module but if you may try on your own machine (not HPC due to access right) if interested.

Evaluate the model by computing the RMSE on the test data:

predictions = model.transform(test)
evaluator = RegressionEvaluator(metricName="rmse", labelCol="rating",predictionCol="prediction")
rmse = evaluator.evaluate(predictions)
print("Root-mean-square error = " + str(rmse))
# Root-mean-square error = 0.9209573052637269

Generate top 10 movie recommendations for each user:

userRecs = model.recommendForAllUsers(10)
userRecs.show(5,  False)
# +------+-------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------+
# |userId|recommendations                                                                                                                                                                      |
# +------+-------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------+
# |471   |[[1643, 6.5805216], [1605, 5.3627353], [309, 5.275607], [613, 5.146943], [114, 5.0345564], [133, 4.8838453], [1159, 4.8557506], [869, 4.833519], [1302, 4.8330326], [776, 4.8192673]]|
# |463   |[[113, 4.39458], [6, 4.17109], [169, 4.030159], [1005, 3.98091], [1252, 3.9767776], [867, 3.9396996], [516, 3.9310303], [1449, 3.919868], [246, 3.908695], [915, 3.9079194]]         |
# |833   |[[1368, 4.767944], [1021, 4.528815], [1597, 4.501705], [340, 4.4845786], [1463, 4.4715357], [1524, 4.451138], [56, 4.3978567], [1128, 4.372282], [634, 4.359913], [646, 4.26006]]    |
# |496   |[[1589, 4.4875054], [114, 4.471943], [320, 4.4689455], [1240, 4.3582573], [119, 4.3407416], [1085, 4.2807164], [613, 4.2485986], [1643, 4.22283], [475, 4.216228], [390, 4.1804137]] |
# |148   |[[50, 4.7727013], [1664, 4.7229643], [1463, 4.719222], [173, 4.6712356], [169, 4.6699014], [168, 4.635963], [115, 4.607701], [478, 4.5947466], [1367, 4.5931487], [172, 4.5518513]]  |
# +------+-------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------+
# only showing top 5 rows

Generate top 10 user recommendations for each movie

movieRecs = model.recommendForAllItems(10)
movieRecs.show(5, False)
# +-------+----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------+
# |movieId|recommendations                                                                                                                                                                   |
# +-------+----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------+
# |1580   |[[38, 1.1592693], [589, 1.0558939], [415, 1.053479], [340, 1.0466049], [580, 1.0436047], [688, 1.0368994], [16, 1.01447], [152, 1.0036571], [475, 0.99923307], [264, 0.9971297]]  |
# |471    |[[341, 4.742011], [628, 4.7299995], [907, 4.711858], [810, 4.7082725], [507, 4.694857], [372, 4.644466], [173, 4.643699], [324, 4.6061506], [688, 4.568334], [849, 4.56669]]      |
# |1591   |[[157, 5.147288], [440, 5.0715895], [681, 4.964819], [861, 4.916547], [810, 4.865007], [212, 4.743086], [696, 4.690889], [443, 4.659895], [697, 4.633425], [512, 4.62862]]        |
# |1342   |[[219, 4.0400357], [565, 3.9486341], [765, 3.9319253], [34, 3.8362799], [662, 3.8200698], [803, 3.8044589], [270, 3.732796], [737, 3.7270598], [449, 3.7244747], [558, 3.6913967]]|
# |463    |[[519, 5.0955424], [928, 5.061705], [440, 4.947975], [810, 4.89811], [173, 4.8721247], [174, 4.761526], [52, 4.744709], [819, 4.7408133], [46, 4.72528], [157, 4.7203035]]        |
# +-------+----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------+
# only showing top 5 rows

Generate top 10 movie recommendations for a specified set of users:

users = ratings.select(als.getUserCol()).distinct().limit(3)
userSubsetRecs = model.recommendForUserSubset(users, 10)
users.show()
# +------+
# |userId|
# +------+
# |    26|
# |    29|
# |   474|
# +------+
userSubsetRecs.show(3,False)
# +------+-------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------+
# |userId|recommendations                                                                                                                                                                      |
# +------+-------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------+
# |26    |[[1449, 4.050502], [884, 3.9505653], [1122, 3.9379153], [1643, 3.93372], [408, 3.9252207], [483, 3.8798523], [318, 3.8776188], [114, 3.8497207], [119, 3.8116813], [127, 3.810518]]  |
# |474   |[[884, 4.9401073], [318, 4.8852944], [1449, 4.8396335], [1122, 4.811585], [408, 4.7511024], [64, 4.7394896], [357, 4.7208233], [1643, 4.7133975], [127, 4.7032666], [1064, 4.697028]]|
# |29    |[[1449, 4.656543], [884, 4.6300607], [963, 4.581265], [272, 4.550117], [408, 4.52602], [114, 4.4871216], [318, 4.4804883], [483, 4.4729433], [64, 4.456217], [1122, 4.4341135]]      |
# +------+-------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------+

Generate top 10 user recommendations for a specified set of movies

movies = ratings.select(als.getItemCol()).distinct().limit(3)
movieSubSetRecs = model.recommendForItemSubset(movies, 10)
movies.show()
# +-------+
# |movieId|
# +-------+
# |    474|
# |     29|
# |     26|
# +-------+
movieSubSetRecs.show(3,False)
# +-------+---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------+
# |movieId|recommendations                                                                                                                                                                  |
# +-------+---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------+
# |26     |[[270, 4.4144955], [341, 4.3995667], [366, 4.3244834], [770, 4.249518], [118, 4.2322574], [414, 4.20005], [274, 4.184001], [923, 4.1715975], [173, 4.168408], [180, 4.1619034]]  |
# |474    |[[928, 5.138258], [810, 5.0860457], [173, 5.0576525], [239, 5.0229735], [794, 4.9939513], [747, 4.9605308], [310, 4.9474325], [686, 4.904606], [339, 4.896157], [118, 4.8948402]]|
# |29     |[[127, 4.425912], [507, 4.315676], [427, 4.2933187], [811, 4.264675], [472, 4.193143], [628, 4.180931], [534, 4.0719366], [907, 4.0206494], [939, 3.993926], [677, 3.967888]]    |
# +-------+---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------+

Let us take a look at the learned factors:

dfItemFactors=model.itemFactors
dfItemFactors.show()
# +---+--------------------+
# | id|            features|
# +---+--------------------+
# | 10|[0.19747244, 0.65...|
# | 20|[0.7627246, 0.401...|
# | 30|[0.6798929, 0.540...|
# | 40|[0.3069924, 0.470...|
# | 50|[-0.14568746, 1.0...|
# | 60|[0.9483578, 1.100...|
# | 70|[0.38704985, 0.29...|
# | 80|[-0.34140116, 1.0...|
# | 90|[0.25251853, 1.20...|
# |100|[0.55230147, 0.68...|
# |110|[0.01884227, 0.20...|
# |120|[0.044651568, 0.1...|
# |130|[-0.22696862, -0....|
# |140|[-0.10749636, 0.1...|
# |150|[0.5518842, 1.098...|
# |160|[0.44457814, 0.80...|
# |170|[0.54556036, 0.77...|
# |180|[0.4252759, 0.709...|
# |190|[0.2943792, 0.781...|
# |200|[0.34967986, 0.44...|
# +---+--------------------+
# only showing top 20 rows

2. Exercises

More movie recommendations

Do the following on HPC. Run your code in batch mode to produce the results.

Download the MovieLens ml-latest-small dataset using wget as in Lab 2 exercises to the ScalableML/Data directory on HPC. Use the unzip command to unzip the files to a directory of your choice (search "unzip linux" to see examples of usage). Read the readme for this dataset to understand the data.
Use ALS to learn five recommendation models on this dataset, using the same split ratio (0.8, 0.2) and seed (6012) as above but five different values of the rank parameter: 5, 10, 15, 20, 25. Plot the (five) resulting RMSE values (on the test set) against the five rank values.
Find the top five movies to recommend to any one user of your choice and display the titles and genres for these five movies (via programming).

3. Additional ideas to explore (optional)

Databricks tutorial

Complete the tasks in the quiz provided by DataBricks on their data or the data from MovieLens directly. A solution is available but you should try before consulting the solution.

Santander Kaggle competition on produce recommendation

A Kaggle competition on Santander Product Recommendation with a prize of USD 60,000, and 1,779 teams participating.
Follow this PySpark notebook on an ALS-based solution.
Learn the way to consider implicit preferences and do the same for other recommendation problems.

Stock Portfolio Recommendations

Follow Chapter ALS: Stock Portfolio Recommendations of PySpark tutorial to perform Stock Portfolio Recommendations)
The data can be downloaded from Online Retail Data Set at UCI.
Pay attention to the data cleaning step that removes rows containing null value. You may need to do the same when you are dealing with real data.
The data manipulation steps are useful to learn.

Context-aware recommendation and time-split evaluation

See the method in Joint interaction with context operation for collaborative filtering and implement it in PySpark.
Perform the time split recommendation as discussed in the paper for the above recommender systems.

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Lab 3 - Scalable matrix factorisation for collaborative filtering.md

Lab 3 - Scalable matrix factorisation for collaborative filtering.md

Lab 3: Matrix Factorisation for Collaborative Filtering in Recommender Systems

Study schedule

Suggested reading

PySpark APIs in pictures (highly recommended)

1. Movie recommendation via collaborative filtering

Get started

Collaborative filtering

Movie recommendation

Explicit vs. implicit feedback

Cold-start problem

MovieLens100k

2. Exercises

More movie recommendations

3. Additional ideas to explore (optional)

Databricks tutorial

Santander Kaggle competition on produce recommendation

Stock Portfolio Recommendations

Context-aware recommendation and time-split evaluation

Files

Lab 3 - Scalable matrix factorisation for collaborative filtering.md

Latest commit

History

Lab 3 - Scalable matrix factorisation for collaborative filtering.md

File metadata and controls

Lab 3: Matrix Factorisation for Collaborative Filtering in Recommender Systems

Study schedule

Suggested reading

PySpark APIs in pictures (highly recommended)

1. Movie recommendation via collaborative filtering

Get started

Collaborative filtering

Movie recommendation

Explicit vs. implicit feedback

Cold-start problem

MovieLens100k

2. Exercises

More movie recommendations

3. Additional ideas to explore (optional)

Databricks tutorial

Santander Kaggle competition on produce recommendation

Stock Portfolio Recommendations

Context-aware recommendation and time-split evaluation