Evaluations

In building a recommender, it’s common to ask the question, how well does it work?  Ultimately, you’ll only know when you release it live to users and measure it against your targets such as increasing sales or user activity.  It’s unlikely, however, that you’ll get it right first time round and you’ll want to be… Read More Evaluations

Dithering

In this post we’ll look at a relatively low cost but high value technique for improving the quality of your recommendations, named dithering.  It’s a technique that re-orders a list of recommendations by slightly shuffling them around.  It is typically implemented in the recommendation post-processing component.  Despite its value, it doesn’t seem to have been… Read More Dithering