PDR Offers a Recommender System for E-commerce 

Pervasive RushRecommender, a Recommender System for E-commerce, powered by the Pervasive DataRush engine, is designed to help you gain insight to your customer needs and increase customer retention. Actionable information is now the required business practice for superior performance. 

Organizations that employ predictive analytics are emerging with the competitive advantage. While standard reports and key performance indicators (KPI) provide direction, Pervasive DataRush allows organizations to use data strategically by building analytical applications on statistics and predictive data mining. This strategy empowers organizations to move beyond business intelligence and into predictive analytics. 

 

How Would a Recommender System Work for My Business? 

Recommender systems emerged from the need of E-commerce sites to help consumers find products to purchase among an often-overwhelming set of choices. What started as a novelty has turned into a serious business tool:  recommender systems aggregate online customer behavior to find trends and make recommendations based on those trends. Collaborative filtering (CF) is a technique for recommender systems that aims at finding the relationships among a new individual and the existing historical data. The philosophy of collaborative filtering is that users will most likely continue choosing similar products if they have already chosen similar ones.

Although Recommender systems originated from E-commerce sites, they are useful for any organization looking to enhance customer sales.  

How Is It Different? 
 

Pervasive RushRecommender is a scalable and efficient dataflow implementation of collaborative filtering based on weighted co-clustering. Co-clustering is a state-of-the-art data mining algorithm that simultaneously clusters rows and columns. These clusters, co-occurrence of like-minded customers versus products, can be used to predict what other products a customer may have the highest likelihood to purchase. This innovative use of customer latent behavior provides powerful insight and enables cross-selling and up-selling to customers.



Try It Now For FREE!

Download your free trial version of Pervasive RushRecommender supported by Pervasive DataRush 4.2.  Not only does Pervasive DataRush 4.2 bring a set of improvements, but it will perform exceptionally fast providing accurate results in seconds.     

 

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 teams were taking hours to build their model.  We have improved accuracy nearly 22% (8.3% vs. 6.5%) since the challenge.

Benefits:

  • High recommendation accuracy
  • High performance
  • Scalable
  • Full library of custom operators for parallel programming

Additional PDR Solutions:

Technical Reports: