HSMC: Hybrid Sentiment Method for Correlation to Analyze COVID-19 Tweets
Lecture Notes on Data Engineering and Communications Technologies
; 89:991-999, 2022.
Article
in English
| Scopus | ID: covidwho-1620219
ABSTRACT
Along with other catastrophes during covid-19, it is required to grasp the public opinions and reaction to detect how COVID-19 is affecting people emotions? This work proposes a Hybrid Sentiment Method for Correlational (HSMC) analysis to discover and distinguish the people’s opinions toward the recent outbreak by manipulating the English tweets of six countries from January to December 2020. The proposed method’s novelty is an assembling method of a modified Pearson Correlation Coefficient (mPCC) with NRC (National Research Council Canada) Emotion Lexicon dictionary. It engaged four different machine learning algorithms, to measure the HSMC method’s efficiency and compare the accuracy by confusion matrices. The experiments revealed that the NB’s accuracy with HSMC outperformed the LR and peak correlational fear level (27.5%) discovered in the USA tweets, and maximum sadness (20.96%) is detected in Brazilian tweets. © 2022, The Author(s), under exclusive license to Springer Nature Switzerland AG.
Correlational, study; COVID-19, pandemic; Machine, learning, algorithms; Sentiment, classification; Twitter, data, analysis; Learning, algorithms; Machine, learning; Assembling, method; Confusion, matrix; Correlational, analysis; Pearson, correlation, coefficients; Public, opinions; Twitter, data, analyse; Correlation, methods
Full text:
Available
Collection:
Databases of international organizations
Database:
Scopus
Language:
English
Journal:
Lecture Notes on Data Engineering and Communications Technologies
Year:
2022
Document Type:
Article
Similar
MEDLINE
...
LILACS
LIS