Big Data, Data Mining, and Machine Learning: Value Creation for Business Leaders and Practitioners

CHAPTER 9 Recommendation Systems

If we have data, let’s look at data. If all we have are opinions, let’s go with mine.

—Jim Barksdale, former Netscape CEO

WHAT ARE RECOMMENDATION SYSTEMS?

Recommendation systems, also known as recommendation engines, are a type of information system whose purpose is to suggest, or recommend, items or actions to users. The recommendations may consist of retail items (movies, books, etc.) or actions, such as following other users in a social network. Typically, the suggestions are a small subset selected from a large collection, according to criteria such as preferences previously expressed by users. Other possible criteria include age, gender, and location.

The most common strategies followed by recommendation systems are the following.

• Content-based filtering gathers auxiliary information (e.g., user demographics, music genre, keywords, answers to a questionnaire) to generate a profile for each user or item. Users are matched to items based on their profiles. Example: Pandora’s Music Genome Project.
• Collaborative filtering is based on past user behavior. Each user’s rating, purchasing, or viewing history allows the system to establish associations between users with similar behavior and between items of interest to the same users. Example: Netflix.

Collaborative filtering is perhaps the most popular of these strategies, due to its domain-free nature. Among collaborative filtering systems, one can further distinguish between neighborhood-based ...