Demand Sensing Decision Support System
summary
Claims a deployed demand‑sensing system on IBM Cloud that combines a curated data lake, NLP on social and news text, market and public‑health feeds, and client POS to produce hyper‑local forecasts. Core artifacts are the resurgence index and disruption recovery index (county‑level economic risk) plus a consumer safety index that scores fear from social posts hourly; models include LSTM ensembles, a milestone + time‑series ensemble, and an IBM Research multi‑factor attention neural network. Retail pilots report better detection of local spikes (example: weather‑driven lift week 9 missed by exponential smoothing) and deployment at multiple retailers to lower inventory and optimize replenishment.
AI-written summary · about
chapters (10)
- 0:00Demand sensing challenges
- 1:56Demand forecasting approaches
- 3:45Internal and external data
- 5:32Disruption and safety indexes
- 8:08Disruption recovery index
- 11:15Safety index insights
- 15:09IBM Cloud solution
- 18:57Neural network model
- 20:43Enterprise integration
- 21:47Q&A
related talks
43mConsumer Intelligence and the Age of Predictive Marketing Analytics
23mApplying AI and ML in the Current Pandemic World
28mUse case in demand forecasting
48mSmart Retail - Vision, Cognition and Intelligence
28mDemand Forecasting: Accuracy, Stability, Automation
24mPost-Digital Shift: Exponential Business Models for Precision Decisions at Scale
27mOptimizing Omnichannel Order Fulfillment with AI and Advanced Analytics
29mAI & ML in tailored software for decision-making purposes