Building a Recommendation System with R

4.3 (4 reviews total)
By Suresh K. Gorakala , Michele Usuelli
    What do you get with a Packt Subscription?

  • Instant access to this title and 7,500+ eBooks & Videos
  • Constantly updated with 100+ new titles each month
  • Breadth and depth in over 1,000+ technologies

About this book

A recommendation system performs extensive data analysis in order to generate suggestions to its users about what might interest them. R has recently become one of the most popular programming languages for the data analysis. Its structure allows you to interactively explore the data and its modules contain the most cutting-edge techniques thanks to its wide international community. This distinctive feature of the R language makes it a preferred choice for developers who are looking to build recommendation systems.

The book will help you understand how to build recommender systems using R. It starts off by explaining the basics of data mining and machine learning. Next, you will be familiarized with how to build and optimize recommender models using R. Following that, you will be given an overview of the most popular recommendation techniques. Finally, you will learn to implement all the concepts you have learned throughout the book to build a recommender system.

Publication date:
September 2015


Chapter 1. Getting Started with Recommender Systems

How do we buy things in our day-to-day lives? We ask our friends, research the product specifications, compare the product with similar products on the Internet, read the feedback from anonymous users, and then we make decisions. How would it be if there is some mechanism that does all these tasks automatically and recommends the products best suited for you efficiently? A recommender system or recommendation engine is the answer to this question.

In this introductory chapter, we will define a recommender system in terms of the following aspects:

  • Helping to develop an understanding of its definition

  • Explaining its basic functions and providing a general introduction of popular recommender systems

  • Highlighting the importance of evaluation techniques


Understanding recommender systems

Have you ever given a thought to the "People you may know" feature in LinkedIn or Facebook? This feature recommends a list of people whom you might know, who are similar to you based on your friends, friends of friends in your close circle, geographical location, skillsets, groups, liked pages, and so on. These recommendations are specific to you and differ from user to user.

Recommender systems are the software tools and techniques that provide suggestions, such as useful products on e-commerce websites, videos on YouTube, friends' recommendations on Facebook, book recommendations on Amazon, news recommendations on online news websites, and the list goes on.

The main goal of recommender systems is to provide suggestions to online users to make better decisions from many alternatives available over the Web. A better recommender system is directed more towards personalized recommendations by taking into consideration the available digital footprint of the user and information about a product, such as specifications, feedback from the users, comparison with other products, and so on, before making recommendations.


The structure of the book

In this book, we will learn about popular recommender systems that are used the most. We will also look into different machine learning techniques used when building recommendation engines with sample code.

The book is divided into 5 chapters:

  • In Chapter 1, Getting Started with Recommender Systems, you will get a general introduction to recommender systems, such as collaborative filtering recommender systems, content-based recommender systems, knowledge-based recommender systems, and hybrid systems; it will also include a brief definition, real-world examples, and brief details of what one will be learning while building a recommender system.

  • In Chapter 2, Data Mining Techniques Used in Recommender Systems, gives you an overview of different machine learning concepts that are commonly used in building a recommender system and how a data analysis problem can be solved. This chapter includes data preprocessing techniques, such as similarity measures, dimensionality reduction, data mining techniques, and its evaluation techniques. Here similarity measures such as Euclidean distance, Cosine distance, Pearson correlation are explained. We will also cover data mining algorithms such as k-means clustering, support vector machines, decision trees, bagging, boosting, and random forests, along with a popular dimensional reduction technique, PCA. Evaluation techniques such as cross validation, regularization, confusion matrix, and model comparison are explained in brief.

  • In Chapter 3, Recommender Systems, we will discuss collaborative filtering recommender systems, an example for user- and item-based recommender systems, using the recommenderlab R package, and the MovieLens dataset. We will cover model building, which includes exploring data, splitting it into train and test datasets, and dealing with binary ratings. You will have an overview of content-based recommender systems, knowledge-based recommender systems, and hybrid systems.

  • In Chapter 4, Evaluating the Recommender Systems, we will learn about the evaluation techniques for recommender systems, such as setting up the evaluation, evaluating recommender systems, and optimizing the parameters.

  • In Chapter 5, Case Study – Building Your Own Recommendation Engine, we will understand a use case in R, which includes steps such as preparing the data, defining the rating matrix, building a recommender, and evaluating and optimizing a recommender.


Collaborative filtering recommender systems

The basic idea of these systems is that, if two users share the same interests in the past, that is, they liked the same book, they will also have similar tastes in the future. If, for example, user A and user B have a similar purchase history and user A recently bought a book that user B has not yet seen, the basic idea is to propose this book to user B. The book recommendations on Amazon are one good example of this type of recommender system.

In this type of recommendation, filtering items from a large set of alternatives is done collaboratively between users preferences. Such systems are called collaborative filtering recommender systems.

While dealing with collaborative filtering recommender systems, we will learn about the following aspects:

  • How to calculate the similarity between users

  • How to calculate the similarity between items

  • How do we deal with new items and new users whose data is not known

The collaborative filtering approach considers only user preferences and does not take into account the features or contents of the items being recommended. This approach requires a large set of user preferences for more accurate results.


Content-based recommender systems

This system recommends items to users by taking the similarity of items and user profiles into consideration. In simpler terms, the system recommends items similar to those that the user has liked in the past. The similarity of items is calculated based on the features associated with the other compared items and is matched with the user's historical preferences.

As an example, we can assume that, if a user has positively rated a movie that belongs to the action genre, then the system can learn to recommend other movies from the action genre.

While building a content-based recommendation system, we take into consideration the following questions:

  • How do we create similarity between items?

  • How do we create and update user profiles continuously?

This technique doesn't take into consideration the user's neighborhood preferences. Hence, it doesn't require a large user group's preference for items for better recommendation accuracy. It only considers the user's past preferences and the properties/features of the items.


Knowledge-based recommender systems

These types of recommender systems are employed in specific domains where the purchase history of the users is smaller. In such systems, the algorithm takes into consideration the knowledge about the items, such as features, user preferences asked explicitly, and recommendation criteria, before giving recommendations. The accuracy of the model is judged based on how useful the recommended item is to the user. Take, for example, a scenario in which you are building a recommender system that recommends household electronics, such as air conditioners, where most of the users will be first timers. In this case, the system considers features of the items, and user profiles are generated by obtaining additional information from the users, such as specifications, and then recommendations are made. These types of system are called constraint-based recommender systems, which we will learn more about in subsequent chapters.

Before building these types of recommender systems, we take into consideration the following questions:

  • What kind of information about the items is taken into the model?

  • How are user preferences captured explicitly?


Hybrid systems

We build hybrid recommender systems by combining various recommender systems to build a more robust system. By combining various recommender systems, we can eliminate the disadvantages of one system with the advantages of another system and thus build a more robust system. For example, by combining collaborative filtering methods, where the model fails when new items don't have ratings, with content-based systems, where feature information about the items is available, new items can be recommended more accurately and efficiently.

Before building a hybrid model, we consider the following questions:

  • What techniques should be combined to achieve the business solution?

  • How should we combine various techniques and their results for better predictions?


Evaluation techniques

Before rolling out the recommender system to the users, how do we ensure that the system is efficient or accurate? What is the base on which we state that the system is good? As stated earlier, the goal of any recommendation system is to recommend more relevant and useful items to the user. A lot of research has been happening in developing new methods to evaluate the recommender systems to improve the accuracy of the systems.

In Chapter 4, Evaluating the Recommender Systems, we will learn about the different evaluation metrics employed to evaluate the recommender systems, these include setting up the evaluation, evaluating recommender systems, optimizing the parameters. This chapter also focuses on how important evaluating the system is during the design and development phases of building recommender systems and the guidelines to be followed in selecting an algorithm based on the available information about the items and the problem statement. This chapter also covers the different experimental setups in which recommender systems are evaluated.


A case study

In Chapter 5, Case Study – Building Your Own Recommendation Engine, we take a case study and build a recommender system step by step as follows:

  1. We take a real-life case and understand the problem statement and its domain aspects

  2. We then perform the data preparation, data source identification, and data cleansing step

  3. Then, we select an algorithm for the recommender system

  4. We then look into the design and development aspects while building the model

  5. Finally, we evaluate and test the recommender system

The implementation of the recommender system is done using R, and code samples will be provided in the book. At the end of this chapter, you will be confident enough to build your own recommendation engine.


The future scope

In the final chapter, I will wrap up by giving the summary of the book and the topics covered. We will focus on the future scope of the research that you will have to undertake. Then we will provide a brief introduction to the current research topics and advancements happening in the field of recommendation systems. I will also list book references and online resources during the course of this book.



In this chapter, you read a synopsis of the popular recommender systems available on the market. In the next chapter, you will learn about the different machine learning techniques used in recommender systems.

About the Authors

  • Suresh K. Gorakala

    Suresh K. Gorakala is a blogger, data analyst, and consultant on data mining, big data analytics, and visualization tools. Since 2013, he has been writing and maintaining a blog on data science at

    Suresh holds a bachelor's degree in mechanical engineering from SRKR Engineering College, which is affiliated with Andhra University, India.

    He loves generating ideas, building data products, teaching, photography, and travelling. Suresh can be reached at [email protected].You can also follow him on Twitter at @sureshgorakala.

    Browse publications by this author
  • Michele Usuelli

    Michele Usuelli is a data scientist, writer, and R enthusiast specialized in the fields of big data and machine learning. He currently works for Revolution Analytics, the leading R-based company that got acquired by Microsoft in April 2015. Michele graduated in mathematical engineering and has worked with a big data start-up and a big publishing company in the past. He is also the author of R Machine Learning Essentials, Packt Publishing.

    Browse publications by this author

Latest Reviews

(4 reviews total)
Clear and direct. Very happy with my purchase.
Building a Recommendation System with R
Unlock this book and the full library FREE for 7 days
Start now