can genetic algorithm (ga) be used for prediction

Réponses (1)

Walter Roberson
Walter Roberson le 21 Mar 2011

0 votes

You have asked two different questions, one in your title and one in the body.
Can genetic algorithms be used for prediction? Yes. Depending on the problem and on how the processing is done, the success rate could vary from 0% to 100%.
Can genetic algorithms be used to predict a given data set? No. By definition, if a value is already known ahead of time, any calculation involving it is not a prediction.

5 commentaires

thanx walter... actually i m interested in prediction...
so how to proceed
Treat it like a neural network problem: present an input and have it figure out plausible methods, present another input and have it figure out a plausible method for that, etc. Eventually you end up with a set of algorithms and weights that you can feed data through (without feedback) to model plausible scenarios under the conditions represented by the data.
sorry walter to distrub u again .... whts d difference between prediction and forcasting coz in my thsis i have to predict the next location of a mobile host
I don't have an answer to that question. At the moment, there is nothing that comes to mind about how you might use genetic algorithms for that task (I am somewhat distracted with other matters this week.)
no problem walter... thanx a lot for solving my queries .. :)

Connectez-vous pour commenter.

Community Treasure Hunt

Find the treasures in MATLAB Central and discover how the community can help you!

Start Hunting!

Translated by