In this paper we describe the information-gathering problem which can be characterized as transforming large amounts of data obtained from sensors into accurate, concise, timely and meaningful information that can be used by decision makers faced with a specific task and a number of options for performing that task. The approach to this information-gathering problem as described here consists of three phases: data validation, data aggregation and abstraction, and information interpretation. Each of these phases will be described in general, and for each of these phases we describe techniques that are reasonably generic to be applicable in many domains, but domain specific knowledge will of course always be needed too.
- artificial intelligence
- information systems