Pervasive Software's Collaboration with University of Texas at Austin
Recently, Pervasive Software teamed with the University of Texas researchers to build a Pervasive DataRush Recommender System using information theoretic co-clustering for classification.
The implementation of dataflow co-clustering predicted over 100 million ratings in 16.31 minutes and achieved an RMSE of 0.88846. This is an effective real-time prediction runtime of 9.7 μs per rating. Moreover, the implemented co-clustering framework supports a wide variety of other large-scale data mining applications.
This joint project resulted in a scientific paper titled "Pervasive Parallelism in Data Mining: Dataflow Solution to Co-clustering Large and Sparse Netflix Data". This paper was selected for presentation at the 15th Annual SIGKDD 2009 Conference on Knowledge Discovery and Data Mining in Paris, France.
Leading academics and our ground-breaking technology will only yield exciting results for the multicore revolution. Together we are successfully delivering unprecedented applications that will benefit organizations across the globe. The applications we jointly develop can unlock the value of data in massive and compute intensive datasets, which has the potential to impact initiatives in biomedicine, data mining, pharmaceuticals, telecommunications and beyond.
Collaborators
- University of Texas at Austin IDEAL Lab
