Recommender systems might be the most widespread variant of predictive model the average individual might encounter.  

They furnish the basis for recommendations on services like Amazon, Spotify, and Youtube. 

Recommender systems are a very intimidating subject if you’re just a beginner. There is a plethora of data prep techniques, algorithms, and model evaluation strategies. 

Not all of the strategies will be relevant, and as a matter of fact, the state-of-the-art can be ignored for now as you will probably obtain very good outcomes by concentrating on the basics, for example, regard it as a straightforward classification or regression issue. 

It is critical to be aware of the fundamentals and have it all laid out for you in a proper and streamline fashion. For this, our recommendation is to skim through or read through the standard books and papers on the subject and looking at few of the widespread libraries. 

In this guide, you will identify resources you can leverage to begin with recommender systems. 

After going through this guide, you will be aware of: 

• The leading review papers on recommender systems you can leverage to quickly understand the state of the field.  
• The leading books on recommender systems from which you can learn the algorithms and strategies needed when developing and assessing recommender systems. 
• The leading Python libraries and APIs that you can leverage to prototype and develop your own recommender systems. 

Tutorial Summarization 

This tutorial is subdivided into three portions, which are: 

1] Papers on Recommender Systems 

2] Books on Recommender Systems 

3] Recommender Systems Libraries 

Papers on Recommender Systems 

Research papers on recommender systems can assist you very swiftly to get up to speed on the state of the field. 

Particularly, review papers that leverage precise language to define what a recommender system is, the algorithms that can be leveraged, conventional datasets and metrics for contrasting algorithms and hints at the state of the art strategies. 

Through skimming or reading a bunch of research papers on recommender systems, you swiftly develop a foundation from which to delve deeper and begin developing your own systems. 

The domain does not alter that swiftly, and strategies from a decade or two ago will provide you solid outcomes. 

Research papers on recommender systems that are advised to obtain a foundational understanding include: 

• Amazon.com Recommendations: Item-to-item Collaborative Filtering, 2003 
• Matrix Factorization Techniques for Recommender Systems, 2009 
• Recommender Systems, 2012 
• Recommender Systems Survey, 2013 
• Advances in Collaborative Filtering, 2015 

After you have queries about particular techniques, you can identify papers that concentrate on those strategies and delve deeper. 

Literature on Recommender Systems 

Books on recommender systems furnish the space to lay out the field and take you on a tour of the strategies and provide you the detail you require to understand them, with more breadth and detail than a lot shorter review paper. 

Again, provided that the field is quite mature, older literature, like that published ten years ago, should not be immediately ignored. 

Some leading textbooks put out by imminent researchers in the domain include the following: 

• Recommender Systems, An Introduction, 2010 
• Recommender Systems, The Textbook, 2016 

It is advised to obtain a handbook on the subject with chapters authored different academics summarizing or championing their preferred strategies and methods. 

• Recommender Systems Handbook, 2015 
• Practical Recommender Systems 

These are recommended handbooks on the subject. 

Recommender Systems Libraries 

You likely don’t require to dive into the start of the art, at least not to start with. 

As such, standard ML libraries are a great place to begin. 

For instance, you can develop an efficient recommender system leveraging matrix factorization methods (SVD) or even a direct k-nearest neighbours model by items or by users. 

Scikit-learn comes recommended for some experiments to start with. 

You can get up to speed on conventional recommender system datasets if your data is not yet accessible or available, or you just wish to get to a rhythm first. 

Popular standard datasets for recommender systems include: 

• MovieLens 
• Yahoo datasets (music, urls, movies, etc.) 

If you are prepped for state-of-the-art strategies, a great place to begin is “papers with code” that details both academic papers and links to the source code for the strategies detailed in the paper. 

There are a variety of proprietary and open-source libraries and services for recommender systems. 

Open-source Python libraires at the start come recommended, such as: 

• Surprise: A Python scikit for building and analysing recommender systems 
• Case recommender: A flexible and extensible python framework for recommender systems 

Conclusion 

In this guide, you identified resources you can leverage to begin with recommender systems. 

Particularly, you learned: 

• The top review papers on recommender systems you can leverage to swiftly comprehend the state of the domain. 
• The leading books on recommender systems from which you can learn the algorithms and strategies needed when developing and assessing recommender systems. 
• The leading Python libraries and APIs that you can leverage to prototype and generate your own recommender systems.