RSS and Atom have brought some order out of chaos. The information that you subscribe via those protocols magically arrives in your feed reader, but it's still overload when you track several hundred feeds. At DEMOfall 07, mSpoke is giving feed readers more smarts with FeedHub, which formulates individualized RSS feeds with the most relevant posts, based on an "adaptive personalization engine," from a set of feed sources.
"We use advanced topic analysis to look at source feeds and other posts from source feed," said Dave Mawhinney, CEO and founder of mSpoke. From that we determine preferences or rules, which we call 'memes.' Instead of a single algorithm to predict for the masses, we create individualized algorithms and use advanced machine learning to give users a level of transparency and empowerment, recommending the most relevant items."
FeedHub has controls for what ends up in the feed on an individual's feed reader. As individual content preferences change, FeedHub memes evolve and adapt. Users can provide explicit inputs by dragging and dropping items into areas to signify their usefulness. FeedHub's algorithms also capture inputs from other users with similar memes, as well as an individual's del.icio.us or digg selections.
FeedHub is a beta service, with the attendant bugs. The Pittsburgh, PA-based company plans to make money via advertising, but like many of its Web 2.o brethren is first focused on building a great user experience and attracting users, said Sean Ammirati, mSpoke co-founder and vice president of business development and product management.