| Autor: |
Souad, Guessoum, aTayeb, Laskri Mohamed |
| Zdroj: |
2012 IEEE International Conference on Complex Systems (ICCS); 1/ 1/2012, p1-6, 6p |
| Abstrakt: |
Case-Based Reasoning (CBR) is a technique which consists of learning from past experiences. Its use is very interest in domains where experience plays an important role in the resolution of new problems, which is the case in medical diagnosis. This paper presents a decision making support system based on CBR and applied to the diagnosis of Chronic Obstructive Pulmonary Disease (COPD), a dangerous respiratory disease bound to tobacco. In medical activity, the physicians are often in situations where they have to make decision whereas they have not all necessary data then they are essentially based on their experiences to find the most probable diagnosis. Our system aims to reproduce this behavior of physicians by estimating similarity on attributes with missing data in the most important stage of CBR process consisting to retrieve the most similar case. We have proposed implemented and tested three ideas to find the real diagnosis of cases which have missing data. Some heuristics functions have been also developed for measuring similarity on attributes with symbolic nature. Preliminary experimentations of these ideas and heuristics have proved a good impact on results. [ABSTRACT FROM PUBLISHER] |
| Databáze: |
Complementary Index |
| Externí odkaz: |
|
