I am building a system that automatically tells the category of any website, by just putting the URL as an input. Of course, this is a classical classification problem. I was wondering:
1. What input data should I use for classification (what part of the website is most informative of its category - home page content, meta keywords, domain name, a mixture of the above?)
2. What classification algorithm to use, so that the processing time of a new website is minimal. I am considering Bayesian filters (one for each category) but it is not the most computationally effective one( since it will try to evaluate each website as many times as categories there are). Another option I am considering are neural networks, or maybe even SVM.
Any suggestions are welcome!
Hi Steve, I have several devices I use for web analytics's, these are to evaluate various factors throughout the website let me have a bit more info and I will send you a link. What are you wanting to investigate? -
Speed, Age, Back links, Traffic (people visiting the site) Value - the more info you provide the better the tool I will suggest - most of them are cost free. (My website and contact details - http://www.apexbusinesssupport.com/).
- Badges
- Science topic
- Similar topics
- Computer Science and Engineering
- Artificial Intelligence