Ananya Muralidhar
Analysis of RNN Capsule Model for Multiclass Imbalanced Data
Authors: Shreekant Jere, Annapurna P. Patil, Ananya Muralidhar
Publication: Springer, Singapore
Published In: Part of the Lecture Notes on Data Engineering and Communications Technologies book series (LNDECT, volume 93)
Publication Date: 17 January 2022
​
Motivation
In today's digital age, platforms like Twitter have become vital communication hubs, making it imperative to accurately gauge sentiments due to the rise in cyberbullying, misinformation, and the psychological impacts of online interactions. Given the challenges of analyzing sentiment from vast and imbalanced datasets, particularly on social media platforms, the project was motivated by the ambition to refine classification accuracy. Recognizing that traditional models often falter in this domain, we sought to leverage advanced neural networks to provide more precise insights into the emotional undertones of vast digital dialogues.
​
Summary
Sentiment analysis, also known as opinion mining, is an essential process in understanding attitudes, views, and emotions from text. Traditional models can be cumbersome, requiring substantial effort and reliance on linguistic knowledge. This research takes a fresh approach by analyzing a non-linguistic-knowledge RNN-based capsule model, addressing the common challenges presented by multiclass imbalanced emotion datasets in the real world.
​
Keywords: Emotion classification, Imbalanced data, RNN Capsule
​
Key Contributions:
-
Detailed literature on RNN capsule architecture and sentiment classification with imbalanced data is provided.
-
A methodology for implementing the ensemble RNN capsule model utilized to manage the imbalanced data is presented.
-
Experimental analysis is conducted on ensemble RNN capsule model with imbalanced emotion dataset.
​
Ensemble RNN Capsule Model
RNN Capsule Architecture
Models Evaluated: Decision Tree, Random Forest, Naïve Bayes, SVM, RNN Capsule
Citation:
Jere, S., Patil, A.P., Muralidhar, A. (2022). Analysis of RNN Capsule Model for Multiclass Imbalanced Data. In: Karrupusamy, P., Balas, V.E., Shi, Y. (eds) Sustainable Communication Networks and Application. Lecture Notes on Data Engineering and Communications Technologies, vol 93. Springer, Singapore. https://doi.org/10.1007/978-981-16-6605-6_31