Post on 25-Jan-2015
description
A BELIEF RULE BASED (BRB) DECISION SUPPORT SYSTEM TO ASSESS CLINICAL
ASTHMA SUSPICIONMOHAMMAD SHAHADAT HOSSAINA, MD. EMRAN HOSSAINB,
MD. SAIFUDDIN KHALIDC, MOHAMMAD A. HAQUED
A, BDEPARTMENT OF COMPUTER SCIENCE AND ENGINEERING, UNIVERSITY OF CHITTAGONG, BANGLADESHCDEPARTMENT OF LEARNING AND PHILOSOPHY & DDEPARTMENT OF ARCHITECTURE, DESIGN AND MEDIA
TECHNOLOGY, AALBORG UNIVERSITY, DENMARK
SCANDINAVIAN CONFERENCE ON HEALTH INFORMATICS (SHI 2014), 21-22 AUGUST
Aims and Objectives
Asthma & Related Works
Belief Rule Base (BRB)
BRB System to Assess Clinical Asthma Suspicion
Result & Discussion
Conclusion
Presentation Outline
Signs, Symptoms and Uncertainties
Causal relationships between signs and symptoms – representation by If-Then rule
Drawbacks of methodologies and algorithms
Expert system: Belief Rule-Based Inference Methodology Using the Evidential Reasoning (RIMER)
Work-in-progress: Optimal learning model (machine learning – to add experience of experts on real time basis.
Aims and Objectives
Asthma is a common chronic inflammatory disease of the airways characterized by variable and recurring symptoms, reversible airflow obstruction, and bronchospasm.
The most common signs and symptoms are- Cough 2) Breathlessness (Shortest of Breath) 3) Wheeze 4) Chest tightness 5) Respiratory Rate. Existing tools and methods:
Optical breath sensor, proportional logic (PL), first-order logic (FOL) or fuzzy logic (FL), forward chaining and backward chaining for inference engine
Scope: RIMER for a refined knowledge base and an inference mechanism.
Asthma and Related Works
Domain Knowledge Representation using BRB
Belief Rule
Domain Knowledge Representation using BRB (Cont.)
BRB System Prototype
Domain Knowledge Representation using BRB (Cont.)
Inference Procedure
Five input antecedents: cough (A1), breathlessness (A2), wheezing (A3), chest tightness (A4) and respiratory rate (A5).
Three referential values of these antecedent attributes: severe (S), moderate (Mo), mild (M) and normal (N).
Asthma (A6) has (2*4*3*2*2) = 96 belief rules
Domain Knowledge Representation using BRB (Cont.)
The BRBES system architecture
BRBES Interface
Asthma diagnosis by BRBES and expert
The AUC for the BRB system prototype is 0.952 (95% confidence interval = 0.960–1.012), and the AUC for the expert opinion is 0.857 (95% confidence interval = 0.939–1.014).
Results and Discussion (Cont.)
Conclusion
Reduce the medical error and various types of uncertainties Reduce medical cost BRB System employed a novel methodology known as RIMER
allows the handling of various types of uncertainty Currently, an attempt has been undertaken to enhance the
system with the capability to supporting the diagnosis of Asthma Training Module for BRB
THANK YOUSCANDINAVIAN CONFERENCE ON HEALTH INFORMATICS (SHI
2014), 21-22 AUGUST