Actually, it's quite simple. You rate a few books - about 10 of them. To most accurately
predict your tastes, you rate not only books you particularly liked, but also books you didn't like at all. We
then search our data for readers similar to you and when we have found enough who share your taste, we look at
their favourite books. From this we then calculate your reading suggestions and the predictions.
The more books you rate, the more accurate the predictions become.
Then there's a lot of book metadata - like genre. If you like a lot of thrillers, then the algorithm evaluates that for your recommendations. It also takes into account the age of the book or the topics it contains.
If you would like to know more about algorithms and collaborative filtering, please feel free to send us an email.