IDAN-Mediator: A Distributed Environment for Temporal Abstractions

Introduction:

Background: Databases and database technology are playing a critical role in almost all areas of our life. In many domains the data is time-dependent, i.e., each data item represents a fact that was true during a specified interval of time. In the clinical domain the temporal dimension of data is extremely important. Reasoning and maintaining temporal data is an important task that many clinical decision support tools require to implement. IDAN proposes a modular distributed architecture around a Temporal Abstraction Server (T.A.S) which encapsulates this task for any client (decision support tool) that will join the environment. The abstraction method is based on the Knowledge-based Temporal-Abstraction (KBTA) method that was already proposed and implemented in the RESUME system.

IDAN, is a proposed distributed modular web based architecture where a user works directly against a single entity: Temporal Abstraction Server, T.A.S, which encapsulates the complex task of generating a result. Users ask primitive queries (about raw data) or abstract queries transparently. All the queries are about medical standard terms, or about knowledge entities that are defined in the knowledge-base, but do not have a standardized entry. Users open sessions against the T.A.S, the session carries a knowledge identifier and a database identifier. During the session the user sends queries which are then processed against the identified knowledge and database.

The overall IDAN architecture. End users interact with the KNAVE II service to submit time-oriented queries. The temporal mediator, using data from the appropriate local data-source, and temporal-abstraction and visualization knowledge from the appropriate domain-specific knowledge base, answers these queries. The visualization service enables users to visually and dynamically explore the resultant abstractions, using a specialized graphical display. Arrows indicates the “uses” relation. Z-like arrows denote remote connections. KB = knowledge base, DB = Database.

The research is supported by the following projects:

• ALMA is the core component of IDAN architecture. ALMA uses Shahar’s Knowledge-Based Temporal-Abstraction method for reasoning about clinical patient data.
• KNAVE is a software tool which facilitates clinical data analysis, throw visualization, explanation and interactive exploration of large data sets.
• KAT - Temporal Abstraction Knowledge Acquisition Tool.
• Medical Vocabularies Search engines:

1. Loinc search engine
2. ICD/9 search engine
3. CPT search engine

• Temporal Abstraction Visual Query Specification
• Local Clinical Database Terms Converter

Idan Team personal: David Boaz, Maya Galperin, Gil Tahan, Michael Ramati, Denis Klimov and Shay Dolev

Related projects: KNAVE, Medical Vocabularies Search engines, The Converter, KAT, Mediator

IDAN publications
Alma Publications

Back to Research Projects main page