Please log in to watch this conference skillscast.
Collaborative filtering (CF) is a powerful algorithm at the core of many recommender systems. However, it is inherently naïve of features that can further improve the quality of recommendations.
At Elsevier, Adam and Anna have augmented their CF-based research article recommender system with a ‘Learning to Rank’ (LTR) machine learning model that uses a rich array of article features to modify and re-rank recommendations. In addition, this model is constantly adapting to real user feedback, so that recommendation quality improves over time with no manual intervention.
In this talk, we will explore the implementation of the CF algorithm and adaptive LTR model in Apache Spark to produce demonstrably higher quality recommendations to our users, and look at how Spark allows developers and data scientists to work together on a web-scale production system.
YOU MAY ALSO LIKE:
- Embrace the Implicit (SkillsCast recorded in December 2018)
- Web Scraping with GoLang (Online Meetup on 13th August 2020)
- Ceci n’est pas un canard - Machine Learning and Generative Adversarial Networks (SkillsCast recorded in August 2020)
- Digital Discrimination: Cognitive Bias in Machine Learning (SkillsCast recorded in June 2020)
Adaptive Recommender Systems with Apache Spark
Anna is a Senior Data Engineer at Elsevier. She has been a Scala developer for 4 years, working for start-ups before joining the world of research. Anna works on various recommendation systems utilising the latest research in data science and machine learning. She loves all things functional, cats and knitting.
Adam is a Data Engineer at Elsevier, working with data scientists to develop smart recommender systems.