Demographic Inference on Twitter using Recursive Neural Networks

Abstract

In social media, demographic inference is a critical task in order to gain a better understanding of a cohort and to facilitate interacting with one's audience. Most previous work has made independence assumptions over topological, textual and label information on social networks. In this work, we employ recursive neural networks to break down these independence assumptions to obtain inference about demographic characteristics on Twitter. We show that our model performs better than existing models including the state-of-theart.

Rumor detection based on propagation graph neural network with attention mechanism

“…BU-RvNN and TD-RvNN: Two variants of the RvNN model proposed by (Ma, Gao, & Wong, 2018b). Both methods construct the propagation tree following the non-sequential propagation structure of tweets in each post set.…”

“…(Vosoughi, 2017) proposed a system to automatically predict the veracity of rumors, the system identifies rumors based on 3 aspects, namely, linguistic style, characteristics of people involved in propagating information, and network propagation dynamics. To reduce the time and cost of manually designing features, (Ma, Gao, & Wong, 2018b) proposed two recursive neural models for rumor detection. The models automatically extract features by recursively traversing the tree structure in either the top-down or bottom-up manner.…”

“…As time goes on, rumors form its specific propagation structure after being commented and replied. Many authors have worked on rumor detection problem by exploiting the propagation characteristic (Ma, Gao, Wei, Lu, & Wong, 2015;Ma, Gao, & Wong, 2018b). However, these work study the propagation patterns based on the contextual information extracted from posts in chronological order, they only capture the temporal characteristic of rumors.…”

Proceedings of the 2nd International Workshop on Fairness, Accountability, Transparency and Ethics in Multimedia

“…Their source usually comes down to either human annotation or prediction by a computational system. The latter option is magnitudes cheaper and faster, leading to an explosion of research describing how to predict the demographics of Twitter users using everything from regression models built on website traffic data [4] to text analysis of usernames [13] to recursive neural networks [7]. The performance of these models are ultimately validated on a smaller set of Twitter users labeled individually for ethnicity and gender by human annotators, either the researchers themselves or -far more commonly -tens of thousands of anonymous strangers on crowd-sourcing platforms.…”

Caracterização autoral a partir de textos utilizando redes neurais artificiais

“…A predição de gênero e idade são as tarefas de CA mais comuns encontradas na literatura, com diversos estudos que tratam estes problemas a partir de conjuntos de dados que trazem essas informações rotuladas e disponibilizadas publicamente (Guimarães et al, 2017;KIM et al, 2017). Por outro lado, outras tarefas como a predição de variação de idioma (SIERRA et al, 2017), renda (FLEKOVA;PREOTIUC-PIETRO;UNGAR, 2016) ou personalidade (GONZÁLEZ-GALLARDO et al, 2015) são menos comuns, possivelmente porque são informações mais difíceis de obter, e muitas vezes exigindo extensas coletas de dados.…”