The recommendation API finds nearest neighbors on a multi-dimensional data mesh. On the one hand, you can add data to your mesh (e.g. by sending usage data of your website). On the other hand, you can request the nearest neighbor for a certain data profile (e.g. to recommend sub-pages to a user).

Required steps

  • All you have to do is to call the danubeRecommendation method of your DanubeClient instance with your data. With each request, the correlation matrix of your data history is refined and better recommendations are returned.

Optional steps

  • You can call the resetCorrelationMatrix method of your DanubeClient instance to reset the current correlation matrix of your API key. Note that this is a powerful operation as your whole data history will be deleted! If you should accidentally reset your correlation matrix you can contact us and we will try to help.