Dimensional Sentiment Analysis of Korean Text using Data Balancing

2021 ◽  
Vol 48 (7) ◽  
pp. 790-801
Author(s):  
Taehee Jeon ◽  
Changhwan Kim
2021 ◽  
pp. 107755872110352
Author(s):  
Matthew Jura ◽  
Joanne Spetz ◽  
Der-Ming Liou

Job satisfaction is a critical component of the professional work environment and is often ascertained through surveys that include structured or open-ended questions. Using data from 24,543 respondents to California Board of Registered Nursing biennial surveys, this study examines the job satisfaction of registered nurses (RNs) by applying clustering analysis to structured job satisfaction items and sentiment analysis to free-text comments. The clustering analysis identified three job satisfaction groups (low, medium, and high satisfaction). Sentiment analysis scores were significantly associated with the job satisfaction groups in both bivariate and multivariate analyses. Differences between the job satisfaction clusters were mostly driven by satisfaction with workload, adequacy of the clerical support services, adequacy of the number of RN staff, and skills of RN colleagues. In addition, there was dispersion in satisfaction related to involvement in management and policy decisions, recognition for a job well done, and opportunities for professional development.


2018 ◽  
Vol 12 (9) ◽  
pp. 190
Author(s):  
Osama Mohammad Rababah ◽  
Esra F. Alzaghoul ◽  
Hussam N. Fakhouri

With the rapid increase in the size of the data over the internet there is a need for new studies for text data summarization and representation; rather than storing the full text or reading the full text we can store and read a summary that represent the original text. Furthermore, there is a need also to represent the summarized text with visual representation; one picture worth ten thousandwords. In this paper we propose an approach for visual representation of the summarized text;visual resources give creative control over how message is perceived andprovide a faster way to know what where the text about.This approach were implemented and tested on a sample of two datasets one of 50 texts and the other dataset of 80 positive and negative movie comments, the evaluation has been done visually and the percent of success cases has been reported, the precision and recall has been calculated.


Author(s):  
Taweesak Kuhamanee ◽  
Nattaphon Talmongkol ◽  
Krit Chaisuriyakul ◽  
Wimol San-Um ◽  
Noppadon Pongpisuttinun ◽  
...  

Explosion of Web 2.0 had made different social media platforms like Facebook, Twitter, Blogs, etc a data hub for the task of Data Mining. Sentiment Analysis or Opinion mining is an automated process of understanding an opinion expressed by customers. By using Data mining techniques, sentiment analysis helps in determining the polarity (Positive, Negative & Neutral) of views expressed by the end user. Nowadays there are terabytes of data available related to any topic then it can be advertising, politics and Survey Companies, etc. CSAT (Customer Satisfaction) is the key factor for this survey companies. In this paper, we used topic modeling by incorporating a LDA algorithm for finding the topics related to social media. We have used datasets of 900 records for analysis. By analysis, we found three important topics from Survey/Response dataset, which are Customers, Agents & Product/Services. Results depict the CSAT score according to Positive, Negative and Neutral response. We used topic modeling which is a statistical modeling technique. Topic modeling is a technique for categorization of text documents into different topics. This approach helps in better summarization of data according to the topic identification and depiction of polarity classification of sentiments expressed.


2016 ◽  
Vol 40 (2) ◽  
pp. 170-186 ◽  
Author(s):  
Hanjun Lee ◽  
Yongmoo Suh

Purpose – Successful open innovation requires that many ideas be posted by a number of users and that the posted ideas be evaluated to find ideas of high quality. As such, successful open innovation community would have inherently information overload problem. The purpose of this paper is to mitigate the information problem by identifying potential idea launchers, so that they can pay attention to their ideas. Design/methodology/approach – This research chose MyStarbucksIdea.com as a target innovation community where users freely share their ideas and comments. We extracted basic features from idea, comment and user information and added further features obtained from sentiment analysis on ideas and comments. Those features are used to develop classification models to identify potential idea launchers, using data mining techniques such as artificial neural network, decision tree and Bayesian network. Findings – The results show that the number of ideas posted and the number of comments posted are the most significant among the features. And most of comment-related sentiment features found to be meaningful, while most of idea-related sentiment features are not in the prediction of idea launchers. In addition, this study show classification rules for the identification of potential idea launchers. Originality/value – This study dealt with information overload problem in an open innovation context. A large volume of textual customer contents from an innovation community were examined and classification models to mitigate the problem were proposed using sentiment analysis and data mining techniques. Experimental results show that the proposed classification models can help the firm identify potential idea launchers for its efficient business innovation.


Sign in / Sign up

Export Citation Format

Share Document