Hi,
I'm trying to map each url's visited by user in particular time period or location (context ) to topics through topic modelling (I'm planning to use text content of web pages for now later I plan to incorporate meta data and hyperlinks for better modelling). For each particular context I will then get an average topic structure. I then want to use same trained model to map all the url's in user history to corresponding topic structure.I then want to pre download all the web pages that are most similar to current context. For this I need to somehow integrate apache mahout to a browser. I also need to train my model on some server or database and then pipeline it to the client. Please help if you have any suggestions. Regards Aks
