Utility Change Point Detection in Online Social Media: A Revealed Preference Framework

Abstract

This paper deals with change detection of utility maximization behaviour in online social media. Such changes occur due to the effect of marketing, advertising, or changes in ground truth. First, we use the revealed preference framework to detect the unknown time point (change point) at which the utility function changed. We derive necessary and sufficient conditions for detecting the change point. Second, in the presence of noisy measurements, we propose a method to detect the change point and construct a decision test. Also, an optimization criteria is provided to recover the linear perturbation coefficients. Finally, to reduce the computational cost, a dimensionality reduction algorithm using Johnson-Lindenstrauss transform is presented. The results developed are illustrated on two real datasets: Yahoo! Tech Buzz dataset and Youstatanalyzer dataset. By using the results developed in the paper, several useful insights can be gleaned from these data sets. First, the changes in ground truth affecting the utility of the agent can be detected by utility maximization behaviour in online search. Second, the recovered utility functions satisfy the single crossing property indicating strategic substitute behaviour in online search. Third, due to the large number of videos in YouTube, the utility maximization behaviour was verified through the dimensionality reduction algorithm. Finally, using the utility function recovered in the lower dimension, we devise an algorithm to predict total traffic in YouTube.

Dynamics of Information Diffusion and Social Sensing

“…H n can be viewed as the "social message", i.e., personal friends of node n since they directly communicate to node n, while the associated beliefs can be viewed as the "informational message". As described in the remarkable recent paper [28], the social message from personal friends exerts a large social influence 13 -it provides significant incentive (peer pressure) for individual n to comply with the protocol of combining its estimate with π 0 n− and thereby prevent incest. [28] shows that receiving messages from known friends has significantly more influence on an individual than the information in the messages.…”

“…The utility function jump changes at an unknown time instant by a linear perturbation. Given the dataset of probe and responses of an agent, the objective in [13] is to develop a nonparametric test to detect the change point and the utility functions before and after the change, which is henceforth referred to as the change point detection problem.…”

Can A User Guess What Her Followers Want?

“…It typically assumes that each consumer decides to buy a bundle of goods, among several alternatives, on the basis of a (concave) nondecreasing utility function [26,27,4]. The works most closely related to ours [10,32,18,9,7] aim to develop efficient algorithms to estimate utility functions from revealed preference data as well as analyze their sample complexity. However, their problem setting is very different from ours: (i) the utility a consumer obtains from buying a bundle of goods is deterministic, however, the feedback a social media user receives from her followers varies randomly and, thus, the utility a user obtains from sharing a story is stochastic; (ii) a consumer can evaluate the utility of a bundle of goods exactly, however, a social media user needs to guess the utility she will obtain from sharing a story on the basis of an estimation of her followers' preferences from the feedback she received in the past; and, (iii) each consumer's decision is independent, however, each social media user's decision is part of a sequential decision making process.…”

Estimating Rationally Inattentive Utility Functions with Deep Clustering for Framing - Applications in YouTube Engagement Dynamics

“…Users interact on YouTube channels by posting comments and rating videos. Extensive empirical studies [18,22,1,16,15,3] show that comments and ratings from users are influenced by the thumbnail, title, category, and perceived popularity of each video. Here we consider a massive YouTube dataset comprising 6 million videos across 25,000 channels and over a millions users from April 2007 to May 2015.…”