- The post argued that user-centric modeling combines knowledge of the world with knowledge of individual customers to produce richer insights.
- The post said context-aware user models, built from sensor data including location and temporal attributes, leave a more vivid picture of a customer.
- The post said the best way to capture context is to model where data is collected or created, then ensemble and stitch results with historical models.
Bob Dylan* said that.
User-centric modeling is here to stay. Rich insights are available when we combine, knowledge of the world with knowledge of your customer. Yes, one at a time. However, users tangle in a network of events and overlap & become part of each others models. Sensor data can avoid granularity mismatch by building models that are per user. Location, temporal attributes and activities provide context. Context-aware user models therefore leave a more vivid picture of your customer. The best way to capture context is to not. Intelligence in the edge – model where the data is being collected or created gives the best context. Ensembling & stitching them with historical models gives perspective. *I’ll let you be in my dreams if I can be in yours
:quality(75))