First, another poll:
The poll above shows that the biggest barrier to adopting Predictive Analytics (per 43%) is cost.
SAP's Charles Gadalla provided this part of the session.
Predictive Analytics 2.0 combines both Predictive Analysis 1.x and SAP InfiniteInsight as shown above
Predictive Analytics 2.0 offers BW offline support (including the BEx Query), a recommendation engine in one component, APL for HANA SPS9 – sister library to the PAL library – on AFL of HANA
Charles said to look for wide datasets – lots of fields, variables
It uses "Smart BI" with the "light bulb" in Lumira which comes from KXEN
It supports Big data
In the green boxes, under future direction, there will be a unified UI with one seamless experience, supporting massive analytics and huge datasets
APL offer auto algorithms to optimize
Skyrock is aompany in France with 2x increase in acceptance rate using predictions and personalized recommendations
Questions and Answers
Q: What are some key use cases?
A: Churn, loyalty, freemium to premium
Q: Roadmap talked about Hadoop – what is integration with Hadoop?
A: Work closely with distributions for Hadoop – connect to Hadoop, Hive, Spark interface
Q: What is roadmap with integration of R?
A: R is already integrated
Q: What training is available for business analysts to make the right choice
A: tutorials, official training program
Another poll appeared:
Yet another poll:
The business analyst will build models; the above adds up to more than 100% because you could select multiple answers.
There was a demo of predictive analytics using hybris. Embedding the prediction analytics inside your operational applications is interesting.
Meet Charles at ASUG Annual Conference in this session: BI109 Predictive Analytics 2.0: Roadmap, what's new and what's next
Also some upcoming ASUG Predictive Analytics webcasts are listed below: