On 23 April 2013 11:00, Steven Walling steven.walling@gmail.com wrote:
Tracking reader activity on Wikipedia is a _very_ touchy subject for a whole host of legitimate reasons, and to be totally honest I don't think you're going to be able to implement any recommender system based on people's reading habits.
Can't he use *broad* categories for tracking? like instead of registering an activity as by the article or the immediate category in which it fall, we can traverse up the categories until we get a category under which there are atleast 1000 articles. This will ensure that the user is not tracked explicitly, like more or less he might not even notice, that he is being tracked (Note: tracking like this doesnot make us evil :P)
For example: if a reader visits the page [Burmese rupee], then we can traverse up the categories like, Burmese rupee-> Rupee-> Currency denominations-> Currency-> International finance-> International economics-> Economics. And then save the visit as an activity in category of "International economics" or "Economics" rather than in "Rupee" or "Currency".
Though this will make the suggestions less efficient, but it will surely generate insights into the active interests of the user (and thats the purpose). Also, there can always be an option to manually register the article by the user (kindof bookmark/watchlist), and if he does so, the suggestions will improve for that user. . Regards Gaurav