Activity-based Twitter sampling for content-based and user-centric prediction models

Abstract

BackgroundTwitter's public and open nature provides great opportunities for its users to actively participate in sharing their opinions and produce high quality content that is reflective of their tendencies and preferences in their day-to-day life [1]. This vast amount of publicly available user-generated content is applied to many applications ranging from tracking human social behavior [2][3][4], detecting events of interest [5][6][7], to smart business [8] where domain knowledge is collected through social media. These studies are either concerned with pulling Twitter and aggregating tweets as bulk or tracking historical tweets over time in order to find meaningful patterns for targeted events. The main challenge of the former studies is the limitation of the Twitter API in accessing only 1% of all existing tweets. However, despite this limitation, the latter studies are concerned with retrieving historical timelines of users.To tackle the above issues of retrieving more tweets beyond the 1% threshold and obtaining historical timelines, topic-based sampling and REST API are both shown to Abstract Increasingly more applications rely on crowd-sourced data from social media. Some of these applications are concerned with real-time data streams, while others are more focused on acquiring temporal footprints from historical data. Nevertheless, determining the subset of "credible" users is crucial. While the majority of sampling approaches focus on individual static networks, dynamic user activity over time is usually not considered, which may result in activity gaps in the collected data. Models based on noisy and missing data can significantly degrade in performance. In this study, we demonstrate how to sample Twitter users in order to produce more credible data for temporal prediction models. We present an activity-based sampling approach where users are selected based on their historical activities in Twitter. The predictability of the collected content from activity-based and random sampling is compared in a content-based and user-centric temporal model. The results indicate the importance of an activityoriented sampling method for the acquisition of more credible content for temporal models.Keywords: Twitter sampling, Temporal prediction models, Historical timelines, User activity, Activity-based sampling Open Access© The Author(s) 2017. This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. Aghababaei and Makrehchi Hum. Cent. Comput. Inf. Sci. (2017) Page 2 of 20 Aghababaei and Makrehchi Hum. Cent. Comput. Inf. Sci. (2017) 7:3 be more effective [9,10]. In topic-based sampling [11], a set of specific keywords or hashtags are applied to collect tweets through the search API...

Regional Differences in Information Privacy Concerns After the Facebook-Cambridge Analytica Data Scandal

“…Tweepy is a Python library for accessing the standard real-time streaming Twitter API, 2 which allows to freely retrieve tweets that match a given query. If the query is too broad that it includes over 1% of the total number of tweets posted at that time worldwide, the query's response is sampled (Aghababaei and Makrehchi, 2017;Morstatter et al, 2014). The way in which Twitter samples the data is unpublished.…”

Information cascades prediction with attention neural network

“…Feature-based approaches [2,3,[11][12][13][14][15] make the connection between the prediction and various types of hand-crafted features that are extracted from the information cascade, including the structural features of the social network, content features, temporal features, and user features. To predict the popularity of news articles in Yahoo News, Arapakis et al [16] used 10 different features that they extracted from the content of the news articles as well as external sources.…”

“…Having obtained the attended whole structure embedding ṡ and temporal embedding ḣ , we can feed these two embeddings into the inter-gate mechanism to effectively combine these two factors. The proposed inter-gate mechanism can capture the different (11) Fig. 3 Architecture of the Intra-attention Mechanism w.r.t.…”

Practical Methods for Imputing Follower Count Dynamics

“…First, collecting data on all possible Twitter accounts (or any other social media accounts) poses prohibitive storage and bandwidth costs, even absent limitations imposed by the service provider; thus, selection of accounts (including the special case in which a census of a focal account set is attempted) is inevitable. Depending on the method employed (see Aghababaei and Makrehchi 2017), and the manner in which it is implemented, all accounts of interest may not be identified at the same time. This may lead to missingness during the initial observation period.…”