{"id":65127,"date":"2019-07-27T10:14:36","date_gmt":"2019-07-27T04:44:36","guid":{"rendered":"https:\/\/data-flair.training\/blogs\/?p=65127"},"modified":"2024-08-02T15:28:44","modified_gmt":"2024-08-02T09:58:44","slug":"data-science-r-movie-recommendation-system","status":"publish","type":"post","link":"https:\/\/data-flair.training\/blogs\/data-science-r-movie-recommendation-system\/","title":{"rendered":"Machine Learning Project – Data Science Movie Recommendation System Project in R"},"content":{"rendered":"

Have you ever been on an online streaming platform like Netflix, Amazon Prime, Voot? I watched a movie and after some time, that platform started recommending me different movies and TV shows. I wondered, how the movie streaming platform could suggest me content that appealed to me. Then I came across something known as Recommendation System. This system is capable of learning my watching patterns and providing me with relevant suggestions. Having witnessed the fourth industrial revolution where Artificial Intelligence<\/strong><\/em><\/a> and other technologies are dominating the market, I am sure that you must have come across a recommendation system in your everyday life. I am also sure that by this time curiosity must be getting the best of you. Therefore, in this Machine Learning Project, I will teach you to build your own Movie recommendation system. So. let’s start.<\/p>\n

\"R<\/a><\/p>\n

Movie Recommendation System Project using ML<\/h2>\n

The main goal of this machine learning project is to build a recommendation engine that recommends movies to users. This R project is designed to help you understand the functioning of how a recommendation system works. We will be developing an Item Based Collaborative Filter. By the end of this tutorial, you will gain experience of implementing your R, Data Science, and Machine learning skills<\/strong><\/em> in a real-life project.<\/p>\n

Before moving ahead in this movie recommendation system project in ML, you need to know what recommendation system means. Read below to find the answer.<\/p>\n

What is a Recommendation System?<\/h2>\n

A recommendation system provides suggestions to the users through a filtering process that is based on user preferences and browsing history<\/em>. The information about the user is taken as an input. The information is taken from the input that is in the form of browsing data. This information reflects the prior usage of the product as well as the assigned ratings.<\/p>\n

A recommendation system is a platform that provides its users with various contents based on their preferences and likings. A recommendation system takes the information about the user as an input. The recommendation system is an implementation of the machine learning algorithms<\/strong><\/em><\/a>.<\/p>\n

\"recommendation<\/a><\/p>\n

A recommendation system also finds a similarity between the different products. For example, Netflix Recommendation System<\/span><\/strong> provides you with the recommendations of the movies that are similar to the ones that have been watched in the past. Furthermore, there is a collaborative content filtering that provides you with the recommendations in respect with the other users who might have a similar viewing history or preferences.<\/p>\n

There are two types of recommendation systems – Content-Based Recommendation System and Collaborative Filtering Recommendation. In this project of recommendation system in R, we will work on a collaborative filtering recommendation system and more specifically, ITEM based collaborative recommendation system.<\/p>\n

The biggest benefit of the collaborative filtering approach and especially the item-based one is that it can benefit from the interaction between users and items, thus, while recommending items, it can search for patterns in grouping of the items. Thus, this method does not involve comprehending the content of the items, and, therefore, it can be applied to any type of products and services. Also, it is capable of analyzing big data, and the users get their recommendations in real-time.<\/p>\n

You must check how Netflix recommendation engine works<\/a><\/strong><\/em><\/p>\n

How to build a Movie Recommendation System using Machine Learning<\/h2>\n

Dataset<\/h3>\n

In order to build our recommendation system, we have used the MovieLens Dataset. You can find the movies.csv and ratings.csv file that we have used in our Recommendation System Project here<\/strong><\/em><\/a>. This data consists of 105339 ratings applied over 10329 movies.<\/p>\n

Importing Essential Libraries<\/h3>\n

In our Data Science project, we will make use of these four packages – \u2018recommenderlab\u2019, \u2018ggplot2\u2019, \u2018data.table\u2019 and \u2018reshape2\u2019.<\/em><\/p>\n

Code:<\/strong><\/p>\n

library(recommenderlab)<\/pre>\n
Output Screenshot:<\/strong><\/p>\n
\"Data<\/a><\/p>\n
Code:<\/strong><\/p>\n
library(ggplot2)\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0#Author DataFlair\r\nlibrary(data.table)\r\nlibrary(reshape2)<\/pre>\n
Output Screenshot:<\/strong><\/p>\n
\"recommendation<\/a><\/p>\n
Wait! Don’t forget to check our leading guide on<\/strong> R programming classification<\/strong><\/a><\/em><\/p>\n
Retrieving the Data<\/h3>\n
We will now retrieve our data from movies.csv into movie_data dataframe and ratings.csv into rating_data. We will use the str() function to display information about the movie_data dataframe.<\/p>\n
Code:<\/strong><\/p>\n
setwd(\"\/home\/dataflair\/data\/movie_data\")\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0#Author DataFlair\r\nmovie_data <- read.csv(\"movies.csv\",stringsAsFactors=FALSE)\r\nrating_data <- read.csv(\"ratings.csv\")\r\nstr(movie_data)<\/pre>\n
Output Screenshot:<\/strong><\/p>\n
\"ML<\/a><\/p>\n
We can overview the summary of the movies using the summary() function. We will also use the head() function to print the first six lines of movie_data<\/p>\n
Code:<\/strong><\/p>\n
summary(movie_data)\u00a0\u00a0\u00a0\u00a0#Author DataFlair<\/pre>\n
Output Screenshot:<\/strong><\/p>\n
\"Project<\/a><\/p>\n
Code:<\/strong><\/p>\n
head(movie_data)<\/pre>\n
Output Screenshot:<\/strong><\/p>\n
\"head<\/a><\/p>\n
Similarly, we can output the summary as well as the first six lines of the ‘rating_data’ dataframe –<\/p>\n
Code:<\/strong><\/p>\n
summary(rating_data)\u00a0\u00a0\u00a0#Author DataFlair<\/pre>\n
Output Screenshot:<\/strong><\/p>\n
\"Output<\/a><\/p>\n
Code:<\/strong><\/p>\n
head(rating_data)<\/pre>\n
Output Screenshot:<\/strong><\/p>\n
\"rating<\/a><\/p>\n
Revise your R concepts with DataFlair for Free, checkout 120+ FREE R Tutorials<\/a>\u00a0<\/strong><\/em><\/p>\n
Data Pre-processing<\/h3>\n
From the above table, we observe that the userId column, as well as the movieId column, consist of integers. Furthermore, we need to convert the genres present in the movie_data dataframe into a more usable format by the users. In order to do so, we will first create a one-hot encoding to create a matrix that comprises of corresponding genres for each of the films.<\/p>\n
Code:<\/strong><\/p>\n
movie_genre <- as.data.frame(movie_data$genres, stringsAsFactors=FALSE)\r\nlibrary(data.table)\r\nmovie_genre2 <- as.data.frame(tstrsplit(movie_genre[,1], '[|]', \r\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0type.convert=TRUE), \r\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0stringsAsFactors=FALSE) #DataFlair\r\ncolnames(movie_genre2) <- c(1:10)\r\n\r\nlist_genre <- c(\"Action\", \"Adventure\", \"Animation\", \"Children\", \r\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\"Comedy\", \"Crime\",\"Documentary\", \"Drama\", \"Fantasy\",\r\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\"Film-Noir\", \"Horror\", \"Musical\", \"Mystery\",\"Romance\",\r\n\u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0\u00a0\"Sci-Fi\", \"Thriller\", \"War\", \"Western\")\r\ngenre_mat1 <- matrix(0,10330,18)\r\ngenre_mat1[1,] <- list_genre\r\ncolnames(genre_mat1) <- list_genre\r\n\r\nfor (index in 1:nrow(movie_genre2)) {\r\n\u00a0\u00a0for (col in 1:ncol(movie_genre2)) {\r\n\u00a0\u00a0\u00a0\u00a0gen_col = which(genre_mat1[1,] == movie_genre2[index,col]) #Author DataFlair\r\n\u00a0\u00a0\u00a0\u00a0genre_mat1[index+1,gen_col] <- 1\r\n}\r\n}\r\ngenre_mat2 <- as.data.frame(genre_mat1[-1,], stringsAsFactors=FALSE) #remove first row, which was the genre list\r\nfor (col in 1:ncol(genre_mat2)) {\r\n\u00a0\u00a0genre_mat2[,col] <- as.integer(genre_mat2[,col]) #convert from characters to integers\r\n} \r\nstr(genre_mat2)<\/pre>\n
Screenshot:<\/strong><\/p>\n
\"creating<\/a><\/p>\n
Output –<\/strong><\/p>\n
\"R<\/a><\/p>\n
In the next step of Data Pre-processing of R project, we will create a ‘search matrix’ that will allow us to perform an easy search of the films by specifying the genre present in our list.<\/p>\n
Code:<\/strong><\/p>\n
SearchMatrix <- cbind(movie_data[,1:2], genre_mat2[])\r\nhead(SearchMatrix)\u00a0\u00a0\u00a0\u00a0#DataFlair<\/pre>\n
Output Screenshot:<\/strong><\/p>\n
\"search<\/a><\/p>\n
There are movies that have several genres, for example, Toy Story, which is an animated film also falls under the genres of Comedy, Fantasy, and Children. This applies to the majority of the films.<\/p>\n
For our movie recommendation system to make sense of our ratings through recommenderlabs, we have to convert our matrix into a sparse matrix one. This new matrix is of the class \u2018realRatingMatrix\u2019. This is performed as follows:<\/p>\n
Code:<\/strong><\/p>\n
ratingMatrix <- dcast(rating_data, userId~movieId, value.var = \"rating\", na.rm=FALSE)\r\nratingMatrix <- as.matrix(ratingMatrix[,-1]) #remove userIds\r\n#Convert rating matrix into a recommenderlab sparse matrix\r\nratingMatrix <- as(ratingMatrix, \"realRatingMatrix\")\r\nratingMatrix<\/pre>\n
Output Screenshot:<\/strong><\/p>\n
\"matrix<\/a><\/p>\n
Are you facing any trouble in implementing recommendation system project in R? Comment below, DataFlair Team is ready to help you.<\/em><\/p>\n
Let us now overview some of the important parameters that provide us various options for building recommendation systems for movies-<\/p>\n
Code:<\/strong><\/p>\n
recommendation_model <- recommenderRegistry$get_entries(dataType = \"realRatingMatrix\")\r\nnames(recommendation_model)<\/pre>\n
Output Screenshot:<\/strong><\/p>\n
\"Machine<\/a><\/p>\n
Code:<\/strong><\/p>\n
lapply(recommendation_model, \"[[\", \"description\")<\/pre>\n
Output Screenshot:<\/strong><\/p>\n
\"recommendation<\/a><\/p>\n
Want to become the next data scientist? Try out the best way and explore Data Science Tutorials Series<\/a>\u00a0to learn Data Science in an easy way with DataFlair!!<\/strong><\/em><\/p>\n
We will implement a single model in our R project – Item Based Collaborative Filtering.<\/p>\n
Code:<\/strong><\/p>\n
recommendation_model$IBCF_realRatingMatrix$parameters<\/pre>\n
Output Screenshot:<\/strong><\/p>\n
\"implementing<\/a><\/p>\n
Exploring Similar Data<\/h3>\n
Collaborative Filtering involves suggesting movies to the users that are based on collecting preferences from many other users. For example, if a user A likes to watch action films and so does user B, then the movies that the user B will watch in the future will be recommended to A and vice-versa. Therefore, recommending movies is dependent on creating a relationship of similarity between the two users. With the help of recommenderlab, we can compute similarities using various operators like cosine, pearson as well as jaccard.<\/p>\n
Code:<\/strong><\/p>\n
similarity_mat <- similarity(ratingMatrix[1:4, ],\r\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0method = \"cosine\",\r\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0which = \"users\")\r\nas.matrix(similarity_mat)\r\n\r\nimage(as.matrix(similarity_mat), main = \"User's Similarities\")<\/pre>\n
Output Screenshot:<\/strong><\/p>\n
\"collaborative<\/a><\/p>\n
In the above matrix, each row and column represents a user. We have taken four users and each cell in this matrix represents the similarity that is shared between the two users.<\/p>\n
Now, we delineate the similarity that is shared between the films –<\/p>\n
Code:<\/strong><\/p>\n
movie_similarity <- similarity(ratingMatrix[, 1:4], method =\r\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\"cosine\", which = \"items\")\r\nas.matrix(movie_similarity)\r\n\r\nimage(as.matrix(movie_similarity), main = \"Movies similarity\")<\/pre>\n
Output Screenshot:<\/strong><\/p>\n
\"data<\/a><\/p>\n
Let us now extract the most unique ratings –<\/p>\n
rating_values <- as.vector(ratingMatrix@data)\r\nunique(rating_values) # extracting unique ratings<\/pre>\n
Now, we will create a table of ratings that will display the most unique ratings.<\/p>\n
Code:<\/strong><\/p>\n
Table_of_Ratings <- table(rating_values) # creating a count of movie ratings\r\nTable_of_Ratings<\/pre>\n
Output Screenshot:<\/strong><\/p>\n
\"building<\/a><\/p>\n
This is the right time to check your R and Data Science Learning. Try these latest interview questions<\/a> and become a pro.\u00a0<\/strong><\/em><\/p>\n
Most Viewed Movies Visualization<\/h2>\n
In this section of the machine learning project, we will explore the most viewed movies in our dataset. We will first count the number of views in a film and then organize them in a table that would group them in descending order.<\/p>\n
Code:<\/strong><\/p>\n
library(ggplot2)\r\nmovie_views <- colCounts(ratingMatrix) # count views for each movie\r\ntable_views <- data.frame(movie = names(movie_views),\r\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0views = movie_views) # create dataframe of views\r\ntable_views <- table_views[order(table_views$views,\r\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0decreasing = TRUE), ] # sort by number of views\r\ntable_views$title <- NA\r\nfor (index in 1:10325){\r\n\u00a0\u00a0table_views[index,3] <- as.character(subset(movie_data,\r\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0movie_data$movieId == table_views[index,1])$title)\r\n}\r\ntable_views[1:6,]<\/pre>\n
Input Screenshot:<\/strong><\/p>\n
\"R<\/a><\/p>\n
Output –\u00a0<\/strong><\/p>\n
\"movies<\/a><\/p>\n
Now, we will visualize a bar plot for the total number of views of the top films. We will carry this out using ggplot2.<\/p>\n
Code:<\/strong><\/p>\n
ggplot(table_views[1:6, ], aes(x = title, y = views)) +\r\n\u00a0\u00a0geom_bar(stat=\"identity\", fill = 'steelblue') +\r\n\u00a0\u00a0geom_text(aes(label=views), vjust=-0.3, size=3.5) +\r\n\u00a0\u00a0theme(axis.text.x = element_text(angle = 45, hjust = 1)) +\r\n\r\n\u00a0\u00a0ggtitle(\"Total Views of the Top Films\")<\/pre>\n
Input Screenshot:<\/strong><\/p>\n
\"R<\/a><\/p>\n
Output:<\/strong><\/p>\n
\"most<\/a><\/p>\n
From the above bar-plot, we observe that Pulp Fiction is the most-watched film followed by Forrest Gump.<\/p>\n
If you are enjoying this Data Science Recommendation System Project, DataFlair brings another project for you – Credit Card Fraud Detection using R<\/a>. Save the link, you can thank me later<\/strong><\/em>\ud83d\ude09\u00a0<\/strong><\/p>\n
Heatmap of Movie Ratings<\/h2>\n
Now, in this data science project of Recommendation system, we will visualize a heatmap of the movie ratings. This heatmap will contain first 25 rows and 25 columns as follows –<\/p>\n
Code:<\/strong><\/p>\n
image(ratingMatrix[1:20, 1:25], axes = FALSE, main = \"Heatmap of the first 25 rows and 25 columns\")<\/pre>\n
Input Screenshot:<\/strong><\/p>\n
\"visualizing<\/a><\/p>\n
Output:<\/strong><\/p>\n
\"R<\/a><\/p>\n
Performing Data Preparation<\/h2>\n
We will conduct data preparation in the following three steps –<\/p>\n