Text-mining the voice of the people

Evangelopoulos, Nicholas; Visinescu, Lucian L.

doi:10.1145/2076450.2076467

Cited by 42 publications

(19 citation statements)

References 18 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…When governments could listen to, analyze and understand citizen opinions voiced via information communication technologies such as SMS messages and websites, and absorb the feedback into administrative decision-making process [3], operators of NBA teams such as coaching staffs and the general managers could use Twitter to discern players' status when they are not on the basketball court, particularly their moods, and to use the information that is otherwise difficult to obtain to predict their performance in the upcoming games. To this end, text mining and sentiment analysis will be applied on AGC to extract hidden intelligence from unstructured textual content, which allows teams to capture players' moods reflected by their words.…”

Section: Introductionmentioning

confidence: 99%

Measuring NBA Players' Mood by Mining Athlete-Generated Content

2015

2015 48th Hawaii International Conference on System Sciences

View full text Add to dashboard Cite

Online athlete-generated content in social media has high potential to become the information source for both team managers and coaches to discern players' mood status and shaky performance before games. In the existing literature, either in psychology or sport analytics, there is a stream of research that investigated the relationship between athletes' mood and the individual sport performance; however, few of them discussed the causality from the social media perspective. In this study, we look deep into the Athlete-generated content (AGC) and aim to provide a more comprehensive framework to sport operators that incorporates players' social media content into their administrative decision-making process. We obtained a unique and extensive dataset of AGC for active NBA players (in the 2012-13 season) from Twitter and apply sentiment analysis technique to measure the general mood polarity of a player. The general mood was then incorporated into econometrics models to examine its effect on players' individual game performance. The results suggest that the mood of NBA player has significant effect on driving sport performance. This paper explores the possibility of using social media data to measure athletes' mood and predicting the sport performance.

show abstract

Section: Introductionmentioning

confidence: 99%

Measuring NBA Players' Mood by Mining Athlete-Generated Content

2015

2015 48th Hawaii International Conference on System Sciences

View full text Add to dashboard Cite

show abstract

“…An extension of this study will aim to determine the relationships between the words in lyrics and the seasons. In addition, a more detailed form of analysis, such as latent Dirichlet allocation or latent semantic analysis, could be employed to identify the structural patterns in lyrics (Evangelopoulos & Visinescu, ). Furthermore, we focused solely on K‐pop in this study, but our analysis could be extended to identify changes in world music markets.…”

Section: Resultsmentioning

confidence: 99%

Quantitative analysis of a half-century of K-Pop songs: Association rule analysis of lyrics and social network analysis of singers and composers

Yoo

Sohn

2017

Journal of Popular Music Studies

View full text Add to dashboard Cite

The authors conducted statistical data analysis to identify patterns in Korean popular songs and composer–singer networks during the past 50 years. With the use of text mining and association rule analysis, the findings indicate that the words “love” and “heart,” which evoke romantic love, were the most commonly used expressions in K‐pop songs from 1964 to 2012. References to love gradually became more direct over the years of the study, reflecting an increasing willingness to acknowledge and discuss romantic relationships. Additionally, most songs after 2000 included words in English, pointing to K‐pop's growing presence in an international market. Social network analysis indicates that since 1992, the influence of composers has increased compared with that of singers in composer–singer networks.

show abstract

“…Representing a document collection with vectors allows researchers to perform operations such as scoring documents on a query, document classification, as well as document and term clustering [14]. These SVDs can then be rotated to alternatively model the data's behavior and facilitate interpretation in an unsupervised setting as well as labeling in supervised approaches [3,13,15]. Last, post-LSA may include comparing and classifying documents using either cosine similarity technique or by clustering or factor analysis.…”

Section: Visualizing Term Eigenvector Prominence In a Corporate Sociamentioning

confidence: 99%

“…Again, TDM algorithms need Stop Lists to prevent SVDs from considering terms that do not add value. This is done not only to reduce computational complexity, but also to reduce spurious language patterns [3] and to minimize the degree to which the term space is distorted [4].…”

Section: Introductionmentioning

confidence: 99%

Visualizing Term Eigenvector Prominence in a Corporate Social Responsibility Context

Parra

Tremblay

Castellanos

2016

JADI

View full text Add to dashboard Cite

In this study we develop a simplified technique for helping researchers and analysts visualize the alternative prominence of term eigenvectors obtained after exploring term associations (Term Clusters) while conducting Text Data Mining on a collection to Corporate Social Responsibility (CSR) reports. The collection analyzed is comprised of CSR reports produced by 7 US firms (Citi, Coca-Cola, Exxon-Mobil, General Motors, Intel, McDonald’s and Microsoft) in 2004, 2008 and 2012. The analysis is performed by year in order to discern how the prominence of term eigenvectors has evolved for each firm and for different CSR topics. Results indicate that term eigenvectors maintain their prominence when CSR topics are related to the core business of the firm in question.

show abstract

Text-mining the voice of the people

Abstract: Statistical techniques help public leaders turn text in unstructured citizen feedback into responsive e-democracy.

Cited by 42 publications

References 18 publications

Measuring NBA Players' Mood by Mining Athlete-Generated Content

Measuring NBA Players' Mood by Mining Athlete-Generated Content

Quantitative analysis of a half-century of K-Pop songs: Association rule analysis of lyrics and social network analysis of singers and composers

Visualizing Term Eigenvector Prominence in a Corporate Social Responsibility Context

Contact Info

Product

Resources

About