1. Developing of an expert system is extremely difficult, more difficult than creating more conventional software. Good experts are hard to find. Extracting their knowledge is a long, tedious job and coding that knowledge into software is a major chore.
2. Expert systems are expensive. It costs a lot to develop one, test it and deliver it to the end-user.
3. Most expert systems still must be implemented and delivered on a big mainframe or minicomputer. Expert systems can b developed and used on personal computers, of course, but these are smaller, less sophisticated, and often less useful systems. The memory size and speed of a personal computer limits its usefulness.
4. Expert systems are not 100% reliable. Even with the best experts contributing to their design, expert systems aren't perfect or infallible. For that reason their output recommendation must be weighed, tested, and otherwise, scrutinized before it is used. A human being should always provide the final judgement.
There disadvantages however are not impossible to overcome. With further advancements in technological hardware and software, these disadvantages will gradually lessen or disappear.
No comments:
Post a Comment