A Rule Engine for Relevance Assessment in a Contextualized Information Delivery System

摘要

In order to support police officers in their daily activities, we have designed a rule-based system which can deliver contextualized information to police officers, thus supporting decision making. In particular, we present a framework that has been designed on the basis of requirements elicited in a previous study, focusing on the rule language and the engine that essentially defines and allows to configure the behaviour of the system. The rules consist of a body which specifies conditions that need to be fulfilled in a certain context. The head of the rules specifies how the relevance ratings of certain information items for specific users need to be updated given that the conditions in the body are met. On the basis of cumulated ratings, the system generates a user- and context-specific ranking of information items. Quantitative evaluations in terms of precision and recall with respect to a gold standard determined in cooperation with police officers show that the system can cater for the requirements of our end users and yields reasonable precision and recall values.