Redux: friend identification system
Recommendation systems are widely used in e-commerce. By analyzing the behavioral patterns, you can determine with a high accuracy which book or DVD will be of interest to you. Modern stores earn the lion’s share of profits by offering the buyer a product that he didn’t suspect before.
Now the systems of recommendations are reaching a new level of their development, taking root already in human relations. Yesterday, in the form of a beta version, the Redux social network was opened, the participants of which can receive something like the following messages: “The user Tyler was detected. With a probability of 79%, you will want to make contact with him. Here is his dossier and list of interests. ”
Using such a site is like solving a puzzle. You are trying to understand yourself. Who are you and what kind of people do you like. When registering, in addition to standard questions, you will need to pass the Mayer-Briggs test for the indication of psychological types.
')
However, the search for potential friends is carried out not only by the results of the test. As stated, in the heart of Redux some self-learning system works, and all users of the site from time to time receive another batch of questions they need to answer (do you like sports? Do you like teddy bears?). This is the clarifying information that is required by the recommendation engine.
In addition to answering questions, when looking for potential friends, the way people use the site is also verified. And also special significance is attached to similarities in rare interests. The more rare your hobby is, the more likely you will like a person who has the same hobby (this is called the “anomaly filter”).
Startup Redux has already received $ 1.65 million in venture financing.
via TechCrunch
Now the systems of recommendations are reaching a new level of their development, taking root already in human relations. Yesterday, in the form of a beta version, the Redux social network was opened, the participants of which can receive something like the following messages: “The user Tyler was detected. With a probability of 79%, you will want to make contact with him. Here is his dossier and list of interests. ”
Using such a site is like solving a puzzle. You are trying to understand yourself. Who are you and what kind of people do you like. When registering, in addition to standard questions, you will need to pass the Mayer-Briggs test for the indication of psychological types.
')
However, the search for potential friends is carried out not only by the results of the test. As stated, in the heart of Redux some self-learning system works, and all users of the site from time to time receive another batch of questions they need to answer (do you like sports? Do you like teddy bears?). This is the clarifying information that is required by the recommendation engine.
In addition to answering questions, when looking for potential friends, the way people use the site is also verified. And also special significance is attached to similarities in rare interests. The more rare your hobby is, the more likely you will like a person who has the same hobby (this is called the “anomaly filter”).
Startup Redux has already received $ 1.65 million in venture financing.
via TechCrunch
Source: https://habr.com/ru/post/19442/
All Articles