We understand under the reversed user identification the following method: the source of (either a person or an organization) a request does not public his or her or its name or identification number, but the data of the request can be accessed. Based on this data one can send offer for the request, without knowing the requester. This is the reverse of the usual anonym tenders, where the requestor (or announcer) is known, but the offers are anonym. We are interested in reversed user identification, because we develop a system, where the users can record data about their way of life (eating, physical activities, environmental impacts). These data can be accessed by way of life mentors, but it must be a guarantee, that the mentors cannot find out to whom belong the data. If a mentor see that someone has a dangerous way of life (because the owner of the data has overweight and moves to little), then he or she can send to him or her a message. The mentors have to identify themselves in the message.

The system uses the following solutions:

• We use profiles to obtain anonymity. The data of the users are put in clusters by a neuron network. The data of a profile is the average of the data of the users in the cluster. The mentors can see only the data of the profiles.
• If there is a kind of data, which has very few instances, for example who has been eating in the Golden rooster restaurant, then the system generates some non-existing people with this kind of data, but the new instances should not affect the average. This solution gives us a guarantee that if a mentor could see behind a profile, then still he or she has difficulties to find out how is a real person.

Keywords:  reversed user identification, secure protocol