[MAHOUT-877] Enable the parallel ALS recommender to use implicit feedback data

• Type: New Feature
• Status: Closed
• Priority: Major
• Resolution: Fixed
• Affects Version/s: 0.6
• Fix Version/s: 0.6
• Assignee: Sebastian Schelter
• Reporter: Sebastian Schelter
• Created: 08/Nov/11 08:17
• Updated: 31/Jan/24 22:11
• Resolved: 08/Nov/11 10:13

Description

Mahout's ParallelALSFactorizationJob offers a distributed matrix factorization for computing recommendations. The current implementation is only suited for explicit feedback data (ratings) unfortunately.

The majority of usecases has to work with implicit feedback however. The paper "Collaborative Filtering for Implicit Feedback Datasets" http://research.yahoo.com/pub/2433 describes a closely related approach that is aimed at implicit feedback data and should easily be integratable into the current ParallelALSJob.