Pervasive DataRush Recommender 

Pervasive DataRush Recommender helps you gain insight into your customers' needs and increases customer retention.  E-commerce sites can use Pervasive DataRush Recommender to help customers find products to purchase among an often-overwhelming set of choices. By analyzing aggregate online customer behavior to find trends, Pervasive DataRush Recommender can make product recommendations that increase the likelihood of a purchase.

Collaborative filtering (CF) is a technique used by Pervasive that uses the interests of like individuals to make a recommendation for a given individual. The philosophy of collaborative filtering is that users will most likely continue to purchase products if like individuals have already chosen similar ones.

Capabilities

  • Scalable and efficient implementation of collaborative filtering based on the advanced weighted co-clustering algorithm
  • Groups users who share the same rating patterns to calculate a prediction for a given user
  • Innovative use of customer latent behavior provides powerful insight and enables cross-sell and up-sell opportunities

The Netflix Challenge

Running the Netflix challenge data through 20 iterations took only 17 minutes on a commodity 16-core server. That’s 17 minutes to build a complex model of 480,189 users and recommend 17,770  movies!  Other implementations were taking hours to build their model.


Next Steps