WITHDRAWN: A Survey on Event Detection and Prediction Online and Offline Models using Social Media Platforms

Tijare, Poonam; Prathuri, Jhansi Rani

doi:10.1016/j.matpr.2021.02.164

“…BCubed f1: The BCubed f1 of any tweet in the dataset is the average BCubed precision and BCubed recall (see equation (10), (11), (12)). The BCubed precision for a tweet captures the number of tweets in the cluster that have the same category according to the annotated label.…”

Section: Evaluation Methodologymentioning

confidence: 99%

“…More recently, the authors in [11] placed particular emphasis on offline and online event detection approaches in social media. They found that most of the reviewed studies fall under the offline category.…”

Section: Related Workmentioning

confidence: 99%

“…Readers interested in the various taxonomies of the topic detection approaches are referred to [24], [1], [25], [10], [11]. 2) For each approach, examine the existence of any assessment related to the two qualities by answering the following two question (Yes/No):…”

Section: Survey Of Topic Detection Approaches In the Context Of Trans...mentioning

confidence: 99%

“…Lately, researchers have proposed many useful topic detection frameworks from social media for various applications, such as diseases outbreak detection [6], disaster management [7], riots control [8] and, crime and terrorism prevention [9], etc.. These approaches can be generally categorised into various groups, such as online or retrospective and supervised or unsupervised [1], [10], [4], [11]. While the online approaches ( [12], [13], [14]) focus on extracting topics from real-time posts ( i.e.…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Towards a Transparent and an Environmental-Friendly Approach for Short Text Topic Detection: A Review of Methods for Performance, Transparency and Carbon Footprint

Sulaimani¹,

Starkey²

2023

Preprint

1

0

View full text Add to dashboard Cite

<p>Online social media platforms have contributed significantly to the dissemination of user-generated information. Many studies have proposed various techniques to analyse the publicly available short texts to automatically extract topics. The majority of these works have mainly focused on the competitive performance of the proposed approaches. In this paper our main focus is on how to tackle this problem by incorporating two other important qualities: Transparency and Carbon Footprint. These two pillars are cornerstones to fulfil the emerging international demands and to adhere to the new regulations, such as “Right to Explanation” and “Green AI”. Based on these three qualities, this paper compares the most prominent algorithms in this field, such as: Latent Dirichlet Allocation, Non-Negative Matrix Factorization and KMeans as well as two most recent approaches, such as: BERTopic and Contextual Analysis. By using two different corpuses, the methods were evaluated for Performance. On average, the results show that BERTopic is the best performing approach overall in terms of Performance. However, Contextual Analysis achieves the best Performance in one of the two corpuses used. When considering the three qualities together, the results demonstrate the effectiveness and the benefits of the Contextual Analysis method towards a more transparent and a greener approach for the topic detection task.</p>

show abstract

“…These approaches can generally be categorized into various groups, such as online or retrospective, supervised or unsupervised, and with or without neural word embedding 1 (such as word2vec and BERT) [1], [13], [4], [14]. While the online approaches ( [15], [16], [17]) focus on extracting topics from real-time posts (i.e., as soon as posts arrive), the retrospective solutions ( [18], [19]) receive a whole corpus (i.e., collections of recorded posts in the past) as an input for analysis in an offline manner.…”

Section: Introductionmentioning

confidence: 99%

Towards a Transparent and an Environmental-Friendly Approach for Short Text Topic Detection: A Comparison of Methods for Performance, Transparency, and Carbon Footprint

Sulaimani,

Starkey

2023

JAIT

0

View full text Add to dashboard Cite

Online social media platforms have contributed significantly t o t he d issemination o f u ser-generated information. Many studies have proposed various techniques to analyze publicly available short texts to automatically extract topics. The majority of these works have mainly focused on the competitive performance of the proposed approaches. In this paper, our main focus is on how to tackle this problem by incorporating two other important qualities: Transparency and Carbon Footprint. These two pillars are cornerstones to fulfill the emerging international demands and to adhere to the new regulations, such as "Right to Explanation" and "Green AI". Based on these three qualities, this paper compares the most prominent algorithms in this field ( specifically within the category of unsupervised-retrospective learning), such as: Latent Dirichlet Allocation, Non-Negative Matrix Factorization, and K-Means, as well as two most recent approaches, such as: BERTopic and Contextual Analysis. By using two different datasets, the methods were evaluated for Performance. On average, the results show that BERTopic is the best-performing approach overall in terms of Performance. However, Contextual Analysis achieves the best Performance in one of the two datasets used. When considering the three qualities together, the results demonstrate the effectiveness and the benefits of the Contextual Analysis method towards a more transparent and greener approach for the topic detection task.

show abstract

Towards a Transparent and an Environmental-Friendly Approach for Short Text Topic Detection: A Review of Methods for Performance, Transparency and Carbon Footprint

Sulaimani¹,

Starkey²

2023

Preprint

0

View full text Add to dashboard Cite

<p>Online social media platforms have contributed significantly to the dissemination of user-generated information. Many studies have proposed various techniques to analyse the publicly available short texts to automatically extract topics. The majority of these works have mainly focused on the competitive performance of the proposed approaches. In this paper our main focus is on how to tackle this problem by incorporating two other important qualities: Transparency and Carbon Footprint. These two pillars are cornerstones to fulfil the emerging international demands and to adhere to the new regulations, such as “Right to Explanation” and “Green AI”. Based on these three qualities, this paper compares the most prominent algorithms in this field, such as: Latent Dirichlet Allocation, Non-Negative Matrix Factorization and KMeans as well as two most recent approaches, such as: BERTopic and Contextual Analysis. By using two different corpuses, the methods were evaluated for Performance. On average, the results show that BERTopic is the best performing approach overall in terms of Performance. However, Contextual Analysis achieves the best Performance in one of the two corpuses used. When considering the three qualities together, the results demonstrate the effectiveness and the benefits of the Contextual Analysis method towards a more transparent and a greener approach for the topic detection task.</p>

show abstract

WITHDRAWN: A Survey on Event Detection and Prediction Online and Offline Models using Social Media Platforms

Cited by 3 publications

References 39 publications

Towards a Transparent and an Environmental-Friendly Approach for Short Text Topic Detection: A Review of Methods for Performance, Transparency and Carbon Footprint

Towards a Transparent and an Environmental-Friendly Approach for Short Text Topic Detection: A Review of Methods for Performance, Transparency and Carbon Footprint

Towards a Transparent and an Environmental-Friendly Approach for Short Text Topic Detection: A Comparison of Methods for Performance, Transparency, and Carbon Footprint

Towards a Transparent and an Environmental-Friendly Approach for Short Text Topic Detection: A Review of Methods for Performance, Transparency and Carbon Footprint

Contact Info

Product

Resources

About