Big data has exploded onto the scene. The sheer volume and complexity of data requires a delicate balance between data mining and infrastructure technology. That said, not all analytical problems start off complex, and not all users require analytics at scale. Maybe you just need to build a few propensity models across a handful of products – and want to use modern techniques, e.g., Random Forest. SAS has released a new set of High-Performance (HP) Data Mining procedures/nodes that are included in SAS® Enterprise Miner 12.3 (SAS 9.4).
Watch this video to see a short product demo on benefits of using High-Performance nodes in SAS Enterprise Miner. When you use these nodes, the processing is ‘multi-threaded’ across the cores on your existing servers. As your big data needs grow, you can run the exact same procedures/nodes – distributed across many servers (i.e., Hadoop) using SAS® High-Performance Data Mining. You will be able to:
- Run more iterations of your neural network
- Try several model selection options, and
- Include more variables, as compared to running a process using non-HP nodes.
Start using these new nodes today to get the most benefits from your current infrastructure, and when your big data analytics needs appear and require more parallelism, you will be able to use the same data mining process in the future.
Good luck!
Jonathan Wexler
