Cognidium

Derived from ‘cognitio’ {“knowledge” in LATIN}, this tool uses values of attributes to

digà collect à analyse à interpret information available in online resources to generate item & purchase recommendations, which are location & time sensitive.

Over a period of calibrated usage the tool also evolves its selection of attributes values to interpret with greater accuracy using advanced learning techniques.

A continuous evolving predictive analytical engine that combines DIGS data with Transnational & Profile data to generate purchase recommendations. The engine trains itself on recommendations to conversions & continuously improves on its benchmarks.

HIGHLIGHTS:

  • Cognidium is a framework which can be deployed with any data schema to extend information of certain attributes using their native values.
  • Each component of the framework is designed on best practices basis its functionality and to handle high volume of data.
  • Cognidium is developed in the principles of SOA, allowing for vertical & horizontal scaling across & within components to manage for high volumes.
  • Integrated with Proxy Providers, Captcha etc. to Dig though Complex Resources
  • Controller uses proprietary load optimization & queuing algorithms which can scale upwards & downwards basis benchmarked loads.
  • DIGS are independent service utilities, hence as new attributes are identified in the schema, the same can be easily added without any new redeployment.
  • COGNUM leverages a mix of unsupervised, supervised & reinforcement learning algorithms laying over a predictive analytical model to generate a continuous stream of evolving recommendations.