The original paper is in English. Non-English content has been machine-translated and may contain typographical errors or mistranslations. ex. Some numerals are expressed as "XNUMX".
Copyrights notice
The original paper is in English. Non-English content has been machine-translated and may contain typographical errors or mistranslations. Copyrights notice
Les flux en ligne sont diffusés en continu par lots avec des polarités variées et à des moments variables. Le système gérant les flux en ligne doit être formé pour classer toutes les polarités variables se produisant dynamiquement. Le système de classification de polarité conçu pour les flux en ligne doit répondre à deux défis importants : i) stabilité-plasticité, ii) prolifération des catégories. Les défis rencontrés dans la classification de polarité des flux en ligne peuvent être résolus à l'aide de la technique d'apprentissage incrémental, qui sert à apprendre de nouvelles classes de manière dynamique et à conserver également les connaissances acquises précédemment. Cet article propose une nouvelle méthodologie d'apprentissage incrémental, ILOF (Incremental Learning of Online Feeds) pour classer les flux en adoptant des techniques d'apprentissage profond telles que RNN (Recurrent Neural Networks) et LSTM (Long Short Term Memory) ainsi que ELM (Extreme Learning Machine). pour résoudre les problèmes mentionnés ci-dessus. La méthode proposée crée un modèle distinct pour chaque lot à l'aide d'ELM et apprend progressivement des lots formés. L'entraînement de chaque lot évite le recyclage des anciens flux, économisant ainsi du temps d'entraînement et de l'espace mémoire. Les aliments formés peuvent être jetés lorsqu’un nouveau lot d’aliments arrive. Les expériences sont réalisées à l'aide d'ensembles de données standard comprenant des flux longs (IMDB, Sentiment140) et des flux courts (Twitter, WhatsApp et Twitter sur le sentiment des compagnies aériennes) et la méthode proposée a montré des résultats positifs en termes de meilleures performances et de précision.
Suresh JAGANATHAN
Sri Sivasubramaniya Nadar College of Engineering
Sathya MADHUSUDHANAN
Sri Sivasubramaniya Nadar College of Engineering
The copyright of the original papers published on this site belongs to IEICE. Unauthorized use of the original or translated papers is prohibited. See IEICE Provisions on Copyright for details.
Copier
Suresh JAGANATHAN, Sathya MADHUSUDHANAN, "Polarity Classification of Social Media Feeds Using Incremental Learning — A Deep Learning Approach" in IEICE TRANSACTIONS on Fundamentals,
vol. E105-A, no. 3, pp. 584-593, March 2022, doi: 10.1587/transfun.2021EAP1046.
Abstract: Online feeds are streamed continuously in batches with varied polarities at varying times. The system handling the online feeds must be trained to classify all the varying polarities occurring dynamically. The polarity classification system designed for the online feeds must address two significant challenges: i) stability-plasticity, ii) category-proliferation. The challenges faced in the polarity classification of online feeds can be addressed using the technique of incremental learning, which serves to learn new classes dynamically and also retains the previously learned knowledge. This paper proposes a new incremental learning methodology, ILOF (Incremental Learning of Online Feeds) to classify the feeds by adopting Deep Learning Techniques such as RNN (Recurrent Neural Networks) and LSTM (Long Short Term Memory) and also ELM (Extreme Learning Machine) for addressing the above stated problems. The proposed method creates a separate model for each batch using ELM and incrementally learns from the trained batches. The training of each batch avoids the retraining of old feeds, thus saving training time and memory space. The trained feeds can be discarded when new batch of feeds arrives. Experiments are carried out using the standard datasets comprising of long feeds (IMDB, Sentiment140) and short feeds (Twitter, WhatsApp, and Twitter airline sentiment) and the proposed method showed positive results in terms of better performance and accuracy.
URL: https://global.ieice.org/en_transactions/fundamentals/10.1587/transfun.2021EAP1046/_p
Copier
@ARTICLE{e105-a_3_584,
author={Suresh JAGANATHAN, Sathya MADHUSUDHANAN, },
journal={IEICE TRANSACTIONS on Fundamentals},
title={Polarity Classification of Social Media Feeds Using Incremental Learning — A Deep Learning Approach},
year={2022},
volume={E105-A},
number={3},
pages={584-593},
abstract={Online feeds are streamed continuously in batches with varied polarities at varying times. The system handling the online feeds must be trained to classify all the varying polarities occurring dynamically. The polarity classification system designed for the online feeds must address two significant challenges: i) stability-plasticity, ii) category-proliferation. The challenges faced in the polarity classification of online feeds can be addressed using the technique of incremental learning, which serves to learn new classes dynamically and also retains the previously learned knowledge. This paper proposes a new incremental learning methodology, ILOF (Incremental Learning of Online Feeds) to classify the feeds by adopting Deep Learning Techniques such as RNN (Recurrent Neural Networks) and LSTM (Long Short Term Memory) and also ELM (Extreme Learning Machine) for addressing the above stated problems. The proposed method creates a separate model for each batch using ELM and incrementally learns from the trained batches. The training of each batch avoids the retraining of old feeds, thus saving training time and memory space. The trained feeds can be discarded when new batch of feeds arrives. Experiments are carried out using the standard datasets comprising of long feeds (IMDB, Sentiment140) and short feeds (Twitter, WhatsApp, and Twitter airline sentiment) and the proposed method showed positive results in terms of better performance and accuracy.},
keywords={},
doi={10.1587/transfun.2021EAP1046},
ISSN={1745-1337},
month={March},}
Copier
TY - JOUR
TI - Polarity Classification of Social Media Feeds Using Incremental Learning — A Deep Learning Approach
T2 - IEICE TRANSACTIONS on Fundamentals
SP - 584
EP - 593
AU - Suresh JAGANATHAN
AU - Sathya MADHUSUDHANAN
PY - 2022
DO - 10.1587/transfun.2021EAP1046
JO - IEICE TRANSACTIONS on Fundamentals
SN - 1745-1337
VL - E105-A
IS - 3
JA - IEICE TRANSACTIONS on Fundamentals
Y1 - March 2022
AB - Online feeds are streamed continuously in batches with varied polarities at varying times. The system handling the online feeds must be trained to classify all the varying polarities occurring dynamically. The polarity classification system designed for the online feeds must address two significant challenges: i) stability-plasticity, ii) category-proliferation. The challenges faced in the polarity classification of online feeds can be addressed using the technique of incremental learning, which serves to learn new classes dynamically and also retains the previously learned knowledge. This paper proposes a new incremental learning methodology, ILOF (Incremental Learning of Online Feeds) to classify the feeds by adopting Deep Learning Techniques such as RNN (Recurrent Neural Networks) and LSTM (Long Short Term Memory) and also ELM (Extreme Learning Machine) for addressing the above stated problems. The proposed method creates a separate model for each batch using ELM and incrementally learns from the trained batches. The training of each batch avoids the retraining of old feeds, thus saving training time and memory space. The trained feeds can be discarded when new batch of feeds arrives. Experiments are carried out using the standard datasets comprising of long feeds (IMDB, Sentiment140) and short feeds (Twitter, WhatsApp, and Twitter airline sentiment) and the proposed method showed positive results in terms of better performance and accuracy.
ER -