| Automated trouble shooting systems use a rules | | | | much longer that it got burnt and so we do not |
| based expert system for searching through historical | | | | repeat the same mistake. But a rule based instruction |
| knowledge bases in order to arrive at reasons of | | | | tells us that we have to boil the milk for 10 mins and |
| failure based on symptoms. Human reasoning works | | | | then mix 2 tea spoons of sugar etc., Human based |
| by recalling old cases and associating experiential | | | | reasoning tells us that we need to add some more |
| knowledge based on solving similar cases. There are | | | | sugar, or the sugar is enough etc., |
| both advantages and disadvantages of rule based | | | | A disadvantage of rules based or an expert system |
| systems as well as human reasoning based systems. | | | | is that all rules have to be enlisted in order to |
| In a rule based inference engine the description of | | | | retrieve causes, but does not use experiential |
| the problems and the resolution is text based. For | | | | knowledge. A disadvantage of a case based system |
| example the database will contain information on the | | | | is the definition of human reasoning is abstract. |
| slowness of a PC or a crashed hard disk and its | | | | A case based reasoning system works in much the |
| symptoms and possible causes. The next time a | | | | same manner as our memories. We learn new things, |
| similar problem occurs we will be able to query the | | | | we learn more about old things, and we retain and |
| database and retrieve causes, symptoms for failure | | | | reuse new information. It will differ from logic or rules |
| etc., But in most cases human reasoning is not similar. | | | | based engine in aspects of recording and retrieving |
| Rule based deduction and inference is good only for | | | | cases. |
| an artificial world that we have constructed, but | | | | There are many Open source based CBR applications |
| human reasoning is much different and is based on | | | | such as MYCBR which can be built for solving |
| previous cases. We do not however use rules to | | | | computer network problems or car repair problems. |
| infer about a common problem. | | | | There are also several commercial implementations of |
| Let us take the example of cooking a favorite dish, | | | | CBR systems in Medical diagnosis. |
| we are able to recall that we simmered the milk that | | | | |