Evaluation of ML-Based Sentiment Analysis Techniques with Stochastic Gradient Descent and Logistic Regression

Title

Evaluation of ML-Based Sentiment Analysis Techniques with Stochastic Gradient Descent and Logistic Regression

Creator

Goswami M.; Sabata P.

Description

In recent times, along with the expansion of technology, the Internet also has flourished exponentially. World is more connected today not only through the technology, but also through sharing sentiments to express views, either be constructive or destructive in front of the world through social media. Twitter, Facebook, Instagram, etc., are being used as social media to reach the world. The study of understanding peoples emotions, intentions, attitudes from unstructured data is opinion mining/sentiment analysis. This is an application of NLP or text mining. In this paper, an attempt is made to realize sentiment analysis's multiple dimensions using approaches such as ML and NLP-based technqies like word frequency and TF-IDF. Using ML approach, experiments were conducted, and the performance of the predictions was visualized. Three different datasets are used. A comparison of logistic regression (LR) and stochastic gradient descent (SGD) algorithms are compared using two different document representation. An extensive comparison is carried out using three different types of dataset. Amazon instant video datasets, bank dataset and movie reviews datasets are being used for the same. Analysis of performance is accomplished by using different graphs. The results indicate that logistic regression performs better than stochastic gradient descent for movie review dataset by using word frequency and TF-IDF-based approach. 2021, The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd.

Source

Lecture Notes in Electrical Engineering, Vol-740 LNEE, pp. 153-163.

Date

2021-01-01

Publisher

Springer Science and Business Media Deutschland GmbH

Subject

Classification; Machine learning; Opinion mining; Sentiment classification; Stochastic gradient descent; TF-IDF; Word frequency

Coverage

Goswami M., Christ University, Bengaluru, India; Sabata P., Christ University, Bengaluru, India

Rights

Restricted Access

Relation

ISSN: 18761100; ISBN: 978-981336392-2

Format

Online

Language

English

Type

Conference paper

Identifier

https://doi.org/10.1007/978-981-33-6393-9_17

https://www.scopus.com/inward/record.uri?eid=2-s2.0-85105876978&doi=10.1007%2f978-981-33-6393-9_17&partnerID=40&md5=4ad25ae3f89b10aba08f120449579bc8