A user opinion and metadata mining scheme for predicting box office performance of movies in the social network environment

Daehoon Kim, Daeyong Kim, Een Jun Hwang, Hong Gu Choi

Research output: Contribution to journalArticle

5 Citations (Scopus)

Abstract

With the rapid proliferation of social network services (SNS), it has become common for people to express their thoughts or opinions on various subjects, such as political events, movies, or commercial products, using short comments. Though the comments reflect personal opinion or preferences, collectively, these represent public opinion or trends. Mining public opinion or trends from a collection of user comments made on SNS could be very useful for many applications. One interesting application is to predict the box office performance of a new movie from user comments made on the movie's trailer. Such a prediction is, nevertheless, a very complicated task because many factors can have an influence on it. In this paper, we propose a scheme for mining public opinion from a collection of user comments, easily available on social networks, on the trailer of a new movie. Next, we predict whether the movie will be a box office hit, based on public opinion and other properties such as the leading actors, director, and their past works. Through various experiments, we show that our scheme can produce satisfactory results.

Original languageEnglish
Pages (from-to)259-272
Number of pages14
JournalNew Review of Hypermedia and Multimedia
Volume19
Issue number3-4
DOIs
Publication statusPublished - 2013 Dec 1

Fingerprint

Light trailers
Metadata
Experiments

Keywords

  • Box office hit
  • Movie trailer
  • Public opinion
  • Social network
  • Trend

ASJC Scopus subject areas

  • Media Technology
  • Computer Science Applications
  • Information Systems

Cite this

A user opinion and metadata mining scheme for predicting box office performance of movies in the social network environment. / Kim, Daehoon; Kim, Daeyong; Hwang, Een Jun; Choi, Hong Gu.

In: New Review of Hypermedia and Multimedia, Vol. 19, No. 3-4, 01.12.2013, p. 259-272.

Research output: Contribution to journalArticle

@article{729c968db5b546ef9b1d697c4a09a876,
title = "A user opinion and metadata mining scheme for predicting box office performance of movies in the social network environment",
abstract = "With the rapid proliferation of social network services (SNS), it has become common for people to express their thoughts or opinions on various subjects, such as political events, movies, or commercial products, using short comments. Though the comments reflect personal opinion or preferences, collectively, these represent public opinion or trends. Mining public opinion or trends from a collection of user comments made on SNS could be very useful for many applications. One interesting application is to predict the box office performance of a new movie from user comments made on the movie's trailer. Such a prediction is, nevertheless, a very complicated task because many factors can have an influence on it. In this paper, we propose a scheme for mining public opinion from a collection of user comments, easily available on social networks, on the trailer of a new movie. Next, we predict whether the movie will be a box office hit, based on public opinion and other properties such as the leading actors, director, and their past works. Through various experiments, we show that our scheme can produce satisfactory results.",
keywords = "Box office hit, Movie trailer, Public opinion, Social network, Trend",
author = "Daehoon Kim and Daeyong Kim and Hwang, {Een Jun} and Choi, {Hong Gu}",
year = "2013",
month = "12",
day = "1",
doi = "10.1080/13614568.2013.835450",
language = "English",
volume = "19",
pages = "259--272",
journal = "New Review of Hypermedia and Multimedia",
issn = "1361-4568",
publisher = "Taylor and Francis Ltd.",
number = "3-4",

}

TY - JOUR

T1 - A user opinion and metadata mining scheme for predicting box office performance of movies in the social network environment

AU - Kim, Daehoon

AU - Kim, Daeyong

AU - Hwang, Een Jun

AU - Choi, Hong Gu

PY - 2013/12/1

Y1 - 2013/12/1

N2 - With the rapid proliferation of social network services (SNS), it has become common for people to express their thoughts or opinions on various subjects, such as political events, movies, or commercial products, using short comments. Though the comments reflect personal opinion or preferences, collectively, these represent public opinion or trends. Mining public opinion or trends from a collection of user comments made on SNS could be very useful for many applications. One interesting application is to predict the box office performance of a new movie from user comments made on the movie's trailer. Such a prediction is, nevertheless, a very complicated task because many factors can have an influence on it. In this paper, we propose a scheme for mining public opinion from a collection of user comments, easily available on social networks, on the trailer of a new movie. Next, we predict whether the movie will be a box office hit, based on public opinion and other properties such as the leading actors, director, and their past works. Through various experiments, we show that our scheme can produce satisfactory results.

AB - With the rapid proliferation of social network services (SNS), it has become common for people to express their thoughts or opinions on various subjects, such as political events, movies, or commercial products, using short comments. Though the comments reflect personal opinion or preferences, collectively, these represent public opinion or trends. Mining public opinion or trends from a collection of user comments made on SNS could be very useful for many applications. One interesting application is to predict the box office performance of a new movie from user comments made on the movie's trailer. Such a prediction is, nevertheless, a very complicated task because many factors can have an influence on it. In this paper, we propose a scheme for mining public opinion from a collection of user comments, easily available on social networks, on the trailer of a new movie. Next, we predict whether the movie will be a box office hit, based on public opinion and other properties such as the leading actors, director, and their past works. Through various experiments, we show that our scheme can produce satisfactory results.

KW - Box office hit

KW - Movie trailer

KW - Public opinion

KW - Social network

KW - Trend

UR - http://www.scopus.com/inward/record.url?scp=84888624813&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84888624813&partnerID=8YFLogxK

U2 - 10.1080/13614568.2013.835450

DO - 10.1080/13614568.2013.835450

M3 - Article

AN - SCOPUS:84888624813

VL - 19

SP - 259

EP - 272

JO - New Review of Hypermedia and Multimedia

JF - New Review of Hypermedia and Multimedia

SN - 1361-4568

IS - 3-4

ER -