Full Version of Steam's Interactive Recommender Goes Live
The full version of Interactive Recommender, i.e. an interactive advisor who analyses our preferences and recommends games that fit them, has launched on Steam. Over the last several months the project has been tested within Steam Labs.
- Analyzes our games, how much we have played them, and uses this data to recommends further productions;
- Uses machine learning to recognize our preferences;
- Highly configurable.
Some time ago, Valve launched Steam Labs, an initiative that enables us to test new solutions that studio employees experiment with. One of the first was the Interactive Recommender, which was so popular with users that it has just become a permanent part of Steam Labs.
As the name suggests, it is a system of recommendations. It analyzes our Steam Library, checking how much time we've spent playing individual game, and then recommends new titles to us on this basis.
Interactive Recommender has been warmly welcomed by Steam users and has therefore become an integral part of the service. The tool is useful as it uses machine learning to learn about our preferences. As a result, it can give us titles that we would otherwise miss. It's a useful feature as there are so many games in this store that it has become highly difficult, if not outright impossible, to keep up with daily releases.
An additional advantage is the high configurability of the system. We determine whether to look for popular or niche productions, old or new, as well as to use tags to narrow down the lists.
