| Author | Gajjar, Pradeep Subodh |
| Call Number | AIT RSPR no. CS-86-17 |
| Subject(s) | Expert systems (Computer science)
|
| Note | A research study submitted in partial fulfillment of the requirements for the degree of Master of Engineering, School of Engineering and Technology |
| Publisher | Asian Institute of Technology |
| Abstract | First generation expert systems , based on a collection of
case- by -case problem solving heuristics , managed to establish
the fact that programs could equal or even surpass the problem
solving capabilities of human experts , albeit in narrowly
defined domains . However these surf ace systems have been found
to be in adequate for tasks that require an access to a
knowledge of structure and function in the domain simply
because they have no internal representation of it. Some
examples of such tasks are , solving difficult problems where a
resort to such "deeper " knowledge is essential and funning
explanation which stems from an actual understanding of the
domain. This has led researchers to explore means for equipping
expert systems to handle the set ask s which are extremely
essential in areas such as medicine . Two alternatives emerging
from this research are the " compiled " and the " d e e p " models as
the basis of expert reasoning. In the case of the " compiled "
models , rules are arrived at from an actual understanding of
the structure and function of the underlying domain, but the
system it self does not have an explicit representation of it .
Whereas in case of "deep" models , an attempt is made at
representing domain specific knowledge which forms the basis of
reasoning of the expert system. It has ben conjectured that a n
expert system based on a deeper understanding would have the
added benefit of extensibility and knowledge acquisition.
In this study the above mentioned models were reviewed and
the characteristics of "deeper" models were further explored .
An experimental diagnosis system for acute respiratory
Infections ( A R I ) was bui1t which based its reasoning on an
explicit representation of knowledge of the respiratory system.
It was found that using a frame - based representation , the
system developed , was more extensible and was better equipped
at giving explanations of medical significance . |
| Year | 1986 |
| Type | Research Study Project Report (RSPR) |
| School | School of Engineering and Technology (SET) |
| Department | Other Field of Studies (No Department) |
| Academic Program/FoS | Computer Science (CS) |
| Chairperson(s) | Sadananda, Ramakoti ;Kanchana Kanchanasut |
| Examination Committee(s) | Goldin, Sarah E.
|
| Scholarship Donor(s) | Government of Thailand. |
| Degree | Research Studies Project Report (M. Eng.) - Asian Institute of Technology, 1986 |