Approbation of the Behavior Rate Model with Hidden Variables Based on Respondents' Data on Recent Instagram Posts

There are many problems related to the study of human behavior. Behavior rate is a very important characteristic of behavior. However, it is often impossible to obtain a direct assessment of the behavior rate due to resource and capacity constraints, therefore, it is urgent to develop methods for indirect assessment of this characteristic. Previously, a model based on the Bayesian belief network was proposed, which estimates the behavior rate on recent episodes of behavior data, taking into account the fact that respondents may be wrong in their responses due to memory work and other reasons. A questionnaire was developed, which collected the respondents' data about the latest episodes of posts published on the Instagram. Posting on the social network can be considered as a studied behavior. Because the time of post publication is fixed, it is possible to find out the actual behavior rate for the study period. Thus, it became possible to test the behavior rate model with hidden variables on real data. Our research purpose is to test the behavior rate model with hidden variables on real data. Based on the results obtained, we can assume how effective it will be to use this model in tasks related to other behavior types, the rate of which cannot be known using direct methods.