At Cisco, GBS and SW group we are doing a PoC with Watson Open Scale on ICPD on IBM Cloud. Business Idea behind this is, this system can capture the Sudden Sales Spike, and once we catch it, we can prepare it and forecast. LTSM model should be able to forecast in advance for future possible Sudden Spikes when it will happen. This solution will create explanations for Cisco 70k SKUs. Which is a big revenue stream for IBM.
Why is it useful?
|Who would benefit from this IDEA?||The customer will get the explainability of, and it's impact in time-series.|
How should it work?
1- Insight of independent variable impacting (in the hierarchy) the forecast of model or prediction ]. These variables have different times in time-series.
|Priority Justification||This feature is critical for the Cisco PoC success. Without this feature, we won't have the solution Cisco has asked us to PoC. There are two internal Cisco and two other vendors in the competition developing this solution.|
|Customer Name||Cisco Systems|