scholarly journals Improving Feature Representation Based on a Neural Network for Author Profiling in Social Media Texts

2016 ◽  
Vol 2016 ◽  
pp. 1-13 ◽  
Author(s):  
Helena Gómez-Adorno ◽  
Ilia Markov ◽  
Grigori Sidorov ◽  
Juan-Pablo Posadas-Durán ◽  
Miguel A. Sanchez-Perez ◽  
...  
Keyword(s):  
Neural Network ◽  
Social Media ◽  
Data Preprocessing ◽  
Feature Representation ◽  
Social Media Data ◽  
Lexical Resource ◽  
Media Texts ◽  
Author Profiling ◽  
Media Data

We introduce a lexical resource for preprocessing social media data. We show that a neural network-based feature representation is enhanced by using this resource. We conducted experiments on the PAN 2015 and PAN 2016 author profiling corpora and obtained better results when performing the data preprocessing using the developed lexical resource. The resource includes dictionaries of slang words, contractions, abbreviations, and emoticons commonly used in social media. Each of the dictionaries was built for the English, Spanish, Dutch, and Italian languages. The resource is freely available.

scholarly journals Identifying Sleep-Deprived Authors of Tweets: Prospective Study

10.2196/13076 ◽  
2019 ◽  
Vol 6 (12) ◽  
pp. e13076
Author(s):  
Sara Melvin ◽  
Amanda Jamal ◽  
Kaitlyn Hill ◽  
Wei Wang ◽  
Sean D Young
Keyword(s):  
Neural Network ◽  
Social Media ◽  
Sleep Deprivation ◽  
Undergraduate Students ◽  
Area Under The Curve ◽  
The Body ◽  
First Year ◽  
Social Media Data ◽  
Tracking Devices ◽  
Media Data

Background Social media data can be explored as a tool to detect sleep deprivation. First-year undergraduate students in their first quarter were invited to wear sleep-tracking devices (Basis; Intel), allow us to follow them on Twitter, and complete weekly surveys regarding their sleep. Objective This study aimed to determine whether social media data can be used to monitor sleep deprivation. Methods The sleep data obtained from the device were utilized to create a tiredness model that aided in labeling the tweets as sleep deprived or not at the time of posting. Labeled data were used to train and test a gated recurrent unit (GRU) neural network as to whether or not study participants were sleep deprived at the time of posting. Results Results from the GRU neural network suggest that it is possible to classify the sleep-deprivation status of a tweet’s author with an average area under the curve of 0.68. Conclusions It is feasible to use social media to identify students’ sleep deprivation. The results add to the body of research suggesting that social media data should be further explored as a potential source for monitoring health.

Deep recurrent neural network for geographical entities disambiguation on social media data

2019 ◽  
Vol 173 ◽  
pp. 117-127 ◽  
Author(s):  
Cristina Zuheros ◽  
Siham Tabik ◽  
Ana Valdivia ◽  
Eugenio Martínez-Cámara ◽  
Francisco Herrera
Keyword(s):  
Neural Network ◽  
Social Media ◽  
Recurrent Neural Network ◽  
Social Media Data ◽  
Deep Recurrent Neural Network ◽  
Media Data

scholarly journals Identifying Sleep-Deprived Authors of Tweets: Prospective Study (Preprint)

2018 ◽  
Author(s):  
Sara Melvin ◽  
Amanda Jamal ◽  
Kaitlyn Hill ◽  
Wei Wang ◽  
Sean D Young
Keyword(s):  
Neural Network ◽  
Social Media ◽  
Sleep Deprivation ◽  
Undergraduate Students ◽  
Area Under The Curve ◽  
The Body ◽  
First Year ◽  
Social Media Data ◽  
Tracking Devices ◽  
Media Data

BACKGROUND Social media data can be explored as a tool to detect sleep deprivation. First-year undergraduate students in their first quarter were invited to wear sleep-tracking devices (Basis; Intel), allow us to follow them on Twitter, and complete weekly surveys regarding their sleep. OBJECTIVE This study aimed to determine whether social media data can be used to monitor sleep deprivation. METHODS The sleep data obtained from the device were utilized to create a tiredness model that aided in labeling the tweets as sleep deprived or not at the time of posting. Labeled data were used to train and test a gated recurrent unit (GRU) neural network as to whether or not study participants were sleep deprived at the time of posting. RESULTS Results from the GRU neural network suggest that it is possible to classify the sleep-deprivation status of a tweet’s author with an average area under the curve of 0.68. CONCLUSIONS It is feasible to use social media to identify students’ sleep deprivation. The results add to the body of research suggesting that social media data should be further explored as a potential source for monitoring health.

scholarly journals Augmentation of Contextualized Concatenated Word Representation and Dilated Convolution Neural Network for Sentiment Analysis

2021 ◽  
Vol 2021 ◽  
pp. 1-13
Author(s):  
Fazeel Abid ◽  
Ikram Ud Din ◽  
Ahmad Almogren ◽  
Hasan Ali Khattak ◽  
Mirza Waqar Baig
Keyword(s):  
Neural Network ◽  
Social Media ◽  
Sentiment Analysis ◽  
Contextual Information ◽  
Social Media Data ◽  
Dilated Convolution ◽  
Word Representation ◽  
Detail Loss ◽  
Media Data

Deep learning-based methodologies are significant to perform sentiment analysis on social media data. The valuable insights of social media data through sentiment analysis can be employed to develop intelligent applications. Among many networks, convolution neural networks (CNNs) are widely used in many conventional text classification tasks and perform a significant role. However, to capture long-term contextual information and address the detail loss problem, CNNs require stacking multiple convolutional layers. Also, the stacking of convolutional layers has issues requiring massive computations and the tuning of additional parameters. To solve these problems, in this paper, a contextualized concatenated word representation (CCWRs) is initialized from social media data based on text which is essential to misspelled and out of vocabulary words (OOV). In CCWRs, different word representation models, for example, Word2Vec, its optimized version FastText and Global Vectors, and GloVe, collectively create contextualized representations upon the sequence of input. Second, a three-layered dilated convolutional neural network (3D-CNN) is proposed that places dilated convolution kernels instead of conventional CNN kernels. Incorporating the extension in the receptive field’s size successfully solves the detail loss problem and achieves long-term context information with different dilation rates. Experiments on datasets demonstrate that the proposed framework achieves reliable results with the selection of numerous hyperparameter tuning and configurations for improved optimization leads to reduced computational resources and reliable accuracy.

Sign in / Sign up

Export Citation Format

Share Document