Topical web crawler:Genetic algorithm vs Machine learning based algorithm

See more:

artificial-intelligence

, +

I have decided to implement a web crawler for my CS major project.The project is focused towards adaptive search.I want the pages to be as user specific as possible and time efficiency is not much a constraint.After searching the web for few days(provided I have no prior knowledge in web mining),I realised major research on adaptive web mining reduced down to three major areas:

1)Genetic based algorithms:inspired by evolutionary biology studies.One most researched and efficient algorithm is Infospider<a href="http://cseweb.ucsd.edu/~rik/foa/l2h/foa-7-6-2.html">link</a>.

2)Ant-based algorithms:based on a model of social insect collective behaviour.

3)Machine learning based algorithms:aims at learning stastically characteristics of the linkage structure of the web.For example algorithms based on Hidden Markov model<a href="http://en.wikipedia.org/wiki/Hidden_Markov_model">link</a>.

Now since I have no prior knowledge in web mining.I am unable to decide the most apt algorithm for my project(whether genetic based or machine learning based),given I have 7-8 months to complete my project and I want the project to be as near as possible to practical applications.Also I found both fields(Machine learning and genetic algorithms) to be very intersting.So please enlighten me.