I've read several textbooks but in all of them the definition for heuristics is rather ambiguous and the concept is not explained very well. Could anyone enlighten me? By the way this is in the Evolution option and is a logic paradigm concept.
-
Looking for HSC notes and resources? Check out our Notes & Resources page
Help: Heuristics (1 Viewer)
- Thread starter MarsBarz
- Start date
first try right sub forum http://community.boredofstudies.org/forumdisplay.php?f=284
Then try dictionary
proceeding to a solution by trial and error or by the rules that are only loosely defined - Reader Digest Dictionary
Criteria or principals for deciding which among several alternative courses of action promises to be the most effective in order to achieve some goal. Heuristics are also called rules of thumb... - excel
-Ghost1788
Then try dictionary
proceeding to a solution by trial and error or by the rules that are only loosely defined - Reader Digest Dictionary
Criteria or principals for deciding which among several alternative courses of action promises to be the most effective in order to achieve some goal. Heuristics are also called rules of thumb... - excel
-Ghost1788
i've got Heuristics down as "rules already assumed to be true by expert systems". im not sure if its right because my teacher was helping us create our own definitions and the guy is pretty idiotic. someone confirm it?! btw, whats an example of a heuristic anyway?
The definition I have is
"Heuristic: A rule of thumb often based on experience. It usually results in one or more possible solutions." - Samuel Davis
Heuristics are meant to give the expert system a human feel. e.g. "How are you feeling" has a lot of things to consider (physical health, mental health, personal feeling towards the program etc.) Heuristics would be employed to just get a good solution (an assumption), not always the best solution...
Thats how I understand Heuristics and please correct me if I'm wrong!
"Heuristic: A rule of thumb often based on experience. It usually results in one or more possible solutions." - Samuel Davis
Heuristics are meant to give the expert system a human feel. e.g. "How are you feeling" has a lot of things to consider (physical health, mental health, personal feeling towards the program etc.) Heuristics would be employed to just get a good solution (an assumption), not always the best solution...
Thats how I understand Heuristics and please correct me if I'm wrong!
Heuristics is a rule of thumb and is a way that we humans usually perform a task. For example, we often draw a triangle when doing 3D Trig. So in context, the expert system learns from experience and will perform the task based on experience.
I hope that rights.
I hope that rights.
heuristics helps also lead to the most correct solution (mainly in expert system shells)
for example you might have a set of facts that state what to do if a car isnt starting properly, but if a certain set of facts are true or false, a heuristic may come into place to say "well its actually THIS most likely to be causing it"- before getting to the actual conclusion, its almost like a shortcut.
here is a good example: think of this in real life, if you are experienced with for example fixing a car heaps, and you KNOW that a choking sound is dodgy spark plug, then you instantly diagose that, rather than go through all these steps to check that. - thats what a heuristic is
for example you might have a set of facts that state what to do if a car isnt starting properly, but if a certain set of facts are true or false, a heuristic may come into place to say "well its actually THIS most likely to be causing it"- before getting to the actual conclusion, its almost like a shortcut.
here is a good example: think of this in real life, if you are experienced with for example fixing a car heaps, and you KNOW that a choking sound is dodgy spark plug, then you instantly diagose that, rather than go through all these steps to check that. - thats what a heuristic is
In Sam's textbook, it states that neural networks differ from expert systems because they are able to learn.
So does that mean that expert systems aren't able to "learn" and make assumptions based on prevous experience? Then where do heuristics come in?
This is contadictory. One one hand it says that expert systems aren't able to learn but on the other hand it says that heuristics are rules of thumb based on experience.
Which is right and which is wrong?
This is all on page 303 of the Sam Davis book.
The following statements are contradictory:
" Neural networks differ from expert systems in that they are able to learn."
"Many expert systems also allow the inclusion of heuristics."
"Heuristic - A rule of thumb often based on experience."
I think that I may have overlooked the "often" bit from the last quote. But still, this is ambiguous. In an expert system, what would be the difference between heuristics and standard rules?
So does that mean that expert systems aren't able to "learn" and make assumptions based on prevous experience? Then where do heuristics come in?
This is contadictory. One one hand it says that expert systems aren't able to learn but on the other hand it says that heuristics are rules of thumb based on experience.
Which is right and which is wrong?
This is all on page 303 of the Sam Davis book.
The following statements are contradictory:
" Neural networks differ from expert systems in that they are able to learn."
"Many expert systems also allow the inclusion of heuristics."
"Heuristic - A rule of thumb often based on experience."
I think that I may have overlooked the "often" bit from the last quote. But still, this is ambiguous. In an expert system, what would be the difference between heuristics and standard rules?
neural networks are seperate, they are built purely to learn and my assumption is that heuristics are rules that are learnt, expert systems are purely built with lots of facts (like a database) and able to draw a conclusion from hueristics
you notice how heuristics is like a 'rule' for both, its hard to generally define it, but specifically in AI, expert systems its alot easier to
if you do the 2004 HSC paper there is a question asked similar to what you're saying, look at the answer as well you'll notice CTA uses heurstics as an example for heurstics to draw conclusion with expert system of medical diagnosis
you notice how heuristics is like a 'rule' for both, its hard to generally define it, but specifically in AI, expert systems its alot easier to
if you do the 2004 HSC paper there is a question asked similar to what you're saying, look at the answer as well you'll notice CTA uses heurstics as an example for heurstics to draw conclusion with expert system of medical diagnosis
sikeveo
back after sem2
^ that's my understanding.
sikeveo
back after sem2
Of course. No point denying I am better than you
I haven't come across that term in class yet but have learned it in a Software context to describe an algorithm that seems to perform well most of the time but is only an approximation to make a program faster, for instance using Newtonian physics in a simulation would be a hueristic because it is an approximation of the more correct relativistic physics however this is quite an extreme example seeing how little the difference is.