This course introduces the concepts, applications, algorithms, programming, and design of recommender systems -- software systems that recommend products or information, often based on extensive personalization. Recommender systems are widely used by many users on a day-to-day basis: while recommender systems have been pioneered by e-commerce sites such as Amazon and Netflix, they are widely used ranging in areas from marketing (e.g. personalized product recommendations with your supermarket loyalty card), education, social media, and financial services. This course discusses various techniques for making recommendations, including non-personalized, content-based, and collaborative recommendation techniques, and their evaluation.

TBC

None

 TEXT AS DATA (M) COMPSCI5096

Exam 80%, coursework 20%

Recommender systems are widely deployed stochastic retrieval systems - a search system that learns what to suggest to a user based upon no explicit user input (i.e. zero query), but make suggestions using what is known about the user or the demographics that the user belongs to. With the prevalence of recommender systems, from e-commerce sites to supermarket loyalty cards, it can be seen recommendation technologies have the potential to be more ubiquitous than search technologies. This course introduces the concepts, applications, algorithms, programming, and design of recommender systems -- software systems that recommend products or information, often based on extensive personalization. This course discusses various techniques for making recommendations that are widely used in industry, including non-personalized, content-based, and collaborative recommendation techniques, and their evaluation.

By the end of this course students will be able to:

1. Describe the techniques in making automatic recommendations and personalising them for individual users.

2. Describe different recommendation system scenarios, and models suitable for deployment therein.

3. Explain the theoretical concepts behind different recommendation models.

4. Discuss how a recommender system should be evaluated in terms of the system's performance and the user's satisfaction with the system.

5. Describe the practical engineering issues raised by the implementation of a recommender system in a real deployment.

6. Implement and evaluate a standard recommender system.

7. Discuss leading-edge developments of recommender systems through a critical analysis of a recently published recommender systems paper.

Students must submit at least 75% by weight of the components (including class tests) of the course's summative assessment.