Dataset Information

Rumor Detection over Varying Time Windows.

ABSTRACT: This study determines the major difference between rumors and non-rumors and explores rumor classification performance levels over varying time windows-from the first three days to nearly two months. A comprehensive set of user, structural, linguistic, and temporal features was examined and their relative strength was compared from near-complete date of Twitter. Our contribution is at providing deep insight into the cumulative spreading patterns of rumors over time as well as at tracking the precise changes in predictive powers across rumor features. Statistical analysis finds that structural and temporal features distinguish rumors from non-rumors over a long-term window, yet they are not available during the initial propagation phase. In contrast, user and linguistic features are readily available and act as a good indicator during the initial propagation phase. Based on these findings, we suggest a new rumor classification algorithm that achieves competitive accuracy over both short and long time windows. These findings provide new insights for explaining rumor mechanism theories and for identifying features of early rumor detection.

SUBMITTER: Kwon S

PROVIDER: S-EPMC5230768 | biostudies-literature | 2017

REPOSITORIES: biostudies-literature

ACCESS DATA

Publications

Rumor Detection over Varying Time Windows.

Kwon Sejeong S Cha Meeyoung M Jung Kyomin K

PloS one 20170112 1

This study determines the major difference between rumors and non-rumors and explores rumor classification performance levels over varying time windows-from the first three days to nearly two months. A comprehensive set of user, structural, linguistic, and temporal features was examined and their relative strength was compared from near-complete date of Twitter. Our contribution is at providing deep insight into the cumulative spreading patterns of rumors over time as well as at tracking the pre ...[more]

PMID: 28081135

Similar Datasets

Project description:BackgroundLong-term behavioral and health risk factors constitute a primary focus of research on the etiology of chronic diseases. Yet, identifying critical time-windows during which risk factors have the strongest impact on disease risk is challenging. To assess the trajectory of association of an exposure history with an outcome, the weighted cumulative exposure index (WCIE) has been proposed, with weights reflecting the relative importance of exposures at different times. However, WCIE is restricted to a complete observed error-free exposure whereas exposures are often measured with intermittent missingness and error. Moreover, it rarely explores exposure history that is very distant from the outcome as usually sought in life-course epidemiology.MethodsWe extend the WCIE methodology to (i) exposures that are intermittently measured with error, and (ii) contexts where the exposure time-window precedes the outcome time-window using a landmark approach. First, the individual exposure history up to the landmark time is estimated using a mixed model that handles missing data and error in exposure measurement, and the predicted complete error-free exposure history is derived. Then the WCIE methodology is applied to assess the trajectory of association between the predicted exposure history and the health outcome collected after the landmark time. In our context, the health outcome is a longitudinal marker analyzed using a mixed model.ResultsA simulation study first demonstrates the correct inference obtained with this approach. Then, applied to the Nurses' Health Study (19,415 women) to investigate the association between body mass index history (collected from midlife) and subsequent cognitive decline (evaluated after age 70), the method identified two major critical windows of association: long before the first cognitive evaluation (roughly 24 to 12 years), higher levels of BMI were associated with poorer cognition. In contrast, adjusted for the whole history, higher levels of BMI became associated with better cognition in the last years prior to the first cognitive interview, thus reflecting reverse causation (changes in exposure due to underlying disease).ConclusionsThis approach, easy to implement, provides a flexible tool for studying complex dynamic relationships and identifying critical time windows while accounting for exposure measurement errors.

Project description:Despite their potential importance with regard to tobacco-related health outcomes, as well as their hypothesized role in the production of tobacco-specific N-nitrosamines, bacterial constituents of tobacco products lack characterization. Specifically, to our knowledge, there has been no comprehensive characterization of the effects of storage conditions on the bacterial communities associated with little cigars and cigarillos. To address this knowledge gap, we characterized the bacterial community composition of the tobacco and wrapper components of the following four products: Swisher Sweets Original; Swisher Sweets, Sweet Cherry; Cheyenne Cigars Full Flavor 100's; and Cheyenne Menthol Box. Each product was stored under three different conditions of temperature and relative humidity to mimic different user storage conditions: room (20°C 50% RH), refrigerator (5°C 18% RH) and pocket (25°C 30% RH). On days 0, 5, 9 and 14, subsamples were collected, the wrapper and tobacco were separated, and their total DNA was extracted separately and purified. Resulting DNA was then used in PCR assays targeting the V3 V4 region of the bacterial 16S rRNA gene, followed by sequencing using Illumina HiSeq 300bp PE. Resulting sequences were processed using the Quantitative Insights Into Microbial Ecology (QIIME) software package, followed by analyses in R using the Phyloseq and Vegan packages. A single bacterial phylum, Firmicutes, dominated in the wrapper subsamples whereas the tobacco subsamples were dominated by Proteobacteria. Cheyenne Menthol Box (CMB) samples were characterized by significant differential abundances for 23 bacterial operational taxonomic units (OTUs) in tobacco subsamples and 27 OTUs in the wrapper subsamples between day 0 and day 14 under all conditions. OTUs from the genera Acinetobacter and Bacillus significantly increased in the CMB tobacco subsamples, and OTUs from Bacillus, Streptococcus, Lactobacillus, and Enterococcus significantly increased in the CMB wrapper subsamples over time. These initial results suggest that the bacterial communities of little cigars and cigarillos are dynamic over time and varying storage conditions.

Dataset Information

Rumor Detection over Varying Time Windows.

Publications

Rumor Detection over Varying Time Windows.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets