NORMALIZING USER RESPONSES TO AN EVENT BY REMOVING BEHAVIORAL BIAS

摘要

This document describes techniques for normalizing user responses by removing behavioral bias. In one or more implementations, a response normalization model is built from user responses to one or more events. The response normalization model clusters users into behavioral classification groups based on the user responses. The response normalization model can then be used to normalize user responses by removing behavioral bias from user responses.

主权项

1. A computer-implemented method comprising:
receiving user responses to one or more events from users of a population; determining a response score for each user of the population based on the user responses; clustering the users of the population into behavioral classification groups based on the response scores, the behavioral classification groups including at least a first behavioral classification group that includes users with response scores that are lower relative to the users of the population of users, and a second behavioral classification group that includes users with response scores that are higher relative to the users of the population; and assigning a jump value to each behavioral classification group, the jump value configured to remove behavioral bias from user responses.