An Emotion-Aware Multitask Approach to Fake News and Rumor Detection Using Transfer Learning

Choudhry, Arjun; Khatri, Inder; Jain, Minni; Vishwakarma, Dinesh Kumar

doi:10.1109/tcss.2022.3228312

Cited by 13 publications

(4 citation statements)

References 34 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The instance of data is assumed as 𝐷 and 𝐷 requires the expected outcome to classify the news articles as fake and true which is presented in Eq. (10) |𝐷| denotes the possibility where arbitrary instance 𝐷 belongs to a class 𝐶 𝑖 , the information related to binary encoding is denoted as 𝑙𝑜𝑔 2 and the entropy value is denoted as 𝐸(𝐷). When the attributes present in the data 𝐷 is comprised of distinct values {𝑎 1, 𝑎 2, … , 𝑎 𝑣𝑗, }, then the subset 𝐷 𝑗 it corresponds to 𝑎 𝑗 of 𝐴.…”

Section: ) Id-3 Classifiermentioning

confidence: 99%

See 1 more Smart Citation

An Improved Fake News Detection Model Using Hybrid Time Frequency-Inverse Document Frequency for Feature Extraction and AdaBoost Ensemble Model as a Classifier

Holla,

Kavitha

2024

JAIT

View full text Add to dashboard Cite

Fake news information on the internet has recently emerged as one of the challenging concerns that can have an impact on society and individuals. Furthermore, the propagation of fake news on social media increases the risk of loss of trustworthiness and disseminates fake information across multiple online platforms. As a result, recognizing fake news on the internet plays an important role in society and among individuals. Knowing this, this study proposed an improved fake news detection model that uses the proposed Hybrid Time-Frequency-Inverse Document Frequency (TF-IDF) to extract features and the Adaptive boosting ensemble classifier, which is a combination of Iterative Dichotomiser 3 (ID-3), Random Forest (RF), and Nave Bayes (NB) classifiers. The characteristics are chosen using the Least Absolute Shrinkage and Selection Operator (LASSO), and the news is classified as fake or true using the Adaptive Boosting (AdaBoost) ensemble classifier. The obtained results show that TF-IDF with AdaBoost ensemble classifier has a higher classification accuracy of 98.98% than the existing N-Gram with TF-IDF and Bidirectional Encoder Representation from Transformer (BERT) and Word2Vec with Convolutional Neural Network-Bidirectional Long Short Term Memory (CNN-Bi LSTM) with 96.81% and 97.74%, respectively.

show abstract

Section: ) Id-3 Classifiermentioning

confidence: 99%

“…The fake news detection system helps users to identify and filter illusive information [8,9]. The deceptive nature of fake news creates complications for people to identify the fake and real text (news) [10]. When large scaled news spreads in social media, manual validation becomes difficult.…”

Section: Introductionmentioning

confidence: 99%

An Improved Fake News Detection Model Using Hybrid Time Frequency-Inverse Document Frequency for Feature Extraction and AdaBoost Ensemble Model as a Classifier

Holla,

Kavitha

2024

JAIT

View full text Add to dashboard Cite

show abstract

“…Xu et al (2022) studied the role of comments in rumor detection and proposed a method that extracted the features from the original post and associated comments. Choudhry et al (2022) annotated fake news and rumor datasets with their emotion labels using transfer learning. They proposed a multitasking framework for fake news and rumor detection, predicting the text's emotion and legitimacy.…”

Section: Related Workmentioning

confidence: 99%

An Emotion-Enriched and Psycholinguistics Features-Based Approach for Rumor Detection on Online Social Media

Haque,

Abulaish

2023

Proceedings of the 11th International Workshop on Natural Language Processing for Social Media

View full text Add to dashboard Cite

Social media platforms and online communication tools, once lauded as sources of information, unity, and global connectivity, are now breeding grounds for the unprecedented spread of false, misleading, and manipulative information. Rumor is one type of such information. It consists of untrue information and deceptive messages and is constructed to manipulate emotions. Consequently, emotions are crucial in determining the veracity of rumors. In this paper, we introduce an emotion-enhanced and psycholinguistic features-based approach for rumors detection on social social media. It entails detecting rumors utilizing lexicons and various linguistic features-based learning approaches, primarily by extracting the psychological association of words with their emotions. Emotional and psycholinguistic features are extracted from both posts and comments to enhance the approach and make rumor detection more effective. Using word-level GloVe embedding, the semantic relationships between a post and its comments and their underlying emotions are preserved. The proposed method is evaluated on the popular PHEME dataset and compared to various baselines and SOTA methods, demonstrating substantially superior performance for rumor detection on social media platforms.

show abstract

“…In MTL, the initialisation of the target domain model is usually pre-trained with data from the source domain. MTL is currently used in fault diagnosis and gives good results [19,20]. FTL can change the properties of source and target domains by a domain adaptive method to identify a common potential space.…”

Section: Introductionmentioning

confidence: 99%

A Deep Transfer Learning-Based Network for Diagnosing Minor Faults in the Production of Wireless Chargers

Wang,

Li,

Zhu

2023

Applied Sciences

View full text Add to dashboard Cite

Wireless charger production is critical to energy storage, and effective fault diagnosis of bearings and gears is essential to ensure wireless charging performance with high efficiency, high tolerance to misalignment, and thermal safety. As minor faults are usually difficult to detect, timely diagnosis and detection of minor faults can prevent the fault from worsening and ensure the safety of wireless charging systems. Diagnosing minor faults in bearings and gears with data is a useful but difficult task. To achieve a satisfactory diagnosis of minor faults in the production of wireless charging systems related to the mechanical system that produces wireless charging devices, such as robot arms, this paper proposes a deep learning network based on CNN and LSTM (DTLCL). The method uses deep learning network, model-based transfer learning and range adaptation technology. First, a deep neural network is built to extract significant fault features. Second, the deep transfer network is initialised using model-based transfer learning with a good starting point. Finally, range adaptation using the maximum mean discrepancy between the features learned from the source and target ranges is realised by a multi-layer adaptive technology. The effectiveness of the method was verified using actual measurement data. The training time is 19 s, and the accuracy exceeds 94.5%. The explanation results show that the proposed DTLCL method provides higher accuracy and robust identification of smaller errors compared to the current combination of integrated and single non-transmission models. Due to its data-driven nature, the DTLCL method could be used for fault diagnosis of bearings and gears, which would further promote the application process of wireless charging.

show abstract

An Emotion-Aware Multitask Approach to Fake News and Rumor Detection Using Transfer Learning

Cited by 13 publications

References 34 publications

An Improved Fake News Detection Model Using Hybrid Time Frequency-Inverse Document Frequency for Feature Extraction and AdaBoost Ensemble Model as a Classifier

An Improved Fake News Detection Model Using Hybrid Time Frequency-Inverse Document Frequency for Feature Extraction and AdaBoost Ensemble Model as a Classifier

An Emotion-Enriched and Psycholinguistics Features-Based Approach for Rumor Detection on Online Social Media

A Deep Transfer Learning-Based Network for Diagnosing Minor Faults in the Production of Wireless Chargers

Contact Info

Product

Resources

About