1.6 Context Predictor

How to predict the overall probability of an event, given:

  1. The inputs (each with an observed size)

  2. The Stored parameters on size

  3. The parameters of the importance of each input

 1.6.1 Problem

Of late, banks, retailers, political campaigns, doctors, hospitals and many more enterprises have been quite successful at predicting the behaviour of human beings. Their prediction efforts have been helpful in winning employees, customers and partners. Therefore, predictive context is a key business capability to enable interactive digital experiences.

 1.6.2 Solution

The <ContextPredictor> pattern takes the input queries from a Context-aware application, figures out an appropriate prediction algorithm to answer the queries and have the algorithm deliver its response acting on the data from <InfraContextAggregator>, <DomainContextAggregator> and <AdaptiveSecurity> patterns.

 1.6.3 Application

<ContextPredictor> pattern is about how to solve a prediction problem in enabling context-aware experiences such as the ones we have talked – in such a way a context-aware application can treat it as a black box.

 1.6.4 Examples / Use-cases

What if a healthcare predicts the highest probability “best treatments” based on medical data from hundreds of millions of patients?

What if a heart monitoring system can predict heart attacks hours if not days or years in advance?

What if an IT system in an enterprise can predict a system outage, hours if not days in advance so we can pro-actively prevent it, saving potentially millions of dollars of business impact?