Using neural networks to forecast box office success

Mike Henry, Ramesh Sharda, Dursun Delen

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Abstract

Predicting box office receipts of a particular movie has intrigued many researchers, domain experts and industry leaders as a challenging problem. In this paper, we report on the current status of a prediction system being built at the Institute for Research in Information Systems (IRIS) at Oklahoma State University since 1998. In our model, the forecasting problem is converted into a classification problem, that is, rather than forecasting the pinpoint estimate of box office receipts, a movie is classified into one of nine financial success categories, ranging from a "flop" to a "blockbuster." The prediction results of different datasets representing different time windows and different combination of predictors are presented using average percent hit rate of bingo and oneaway predictions. In the latest tests the prediction results of artificial neural networks improved to almost 50% on "bingo" and close to 90% on "one-away".

Original languageEnglish
Title of host publicationAssociation for Information Systems - 13th Americas Conference on Information Systems, AMCIS 2007
Subtitle of host publicationReaching New Heights
Pages1589-1597
Number of pages9
StatePublished - 1 Dec 2007
Externally publishedYes
Event13th Americas Conference on Information Systems, AMCIS 2007 - Keystone, CO, United States
Duration: 10 Aug 200712 Aug 2007

Publication series

NameAssociation for Information Systems - 13th Americas Conference on Information Systems, AMCIS 2007: Reaching New Heights
Volume3

Conference

Conference13th Americas Conference on Information Systems, AMCIS 2007
CountryUnited States
CityKeystone, CO
Period10/08/0712/08/07

Fingerprint

neural network
Neural networks
movies
information system
expert
leader
industry
Information systems
Industry

Keywords

  • Box-office receipts
  • Classification
  • Data mining
  • Forecasting
  • Motion pictures
  • Neural networks
  • Performance measures
  • Prediction

Cite this

Henry, M., Sharda, R., & Delen, D. (2007). Using neural networks to forecast box office success. In Association for Information Systems - 13th Americas Conference on Information Systems, AMCIS 2007: Reaching New Heights (pp. 1589-1597). (Association for Information Systems - 13th Americas Conference on Information Systems, AMCIS 2007: Reaching New Heights; Vol. 3).
Henry, Mike ; Sharda, Ramesh ; Delen, Dursun. / Using neural networks to forecast box office success. Association for Information Systems - 13th Americas Conference on Information Systems, AMCIS 2007: Reaching New Heights. 2007. pp. 1589-1597 (Association for Information Systems - 13th Americas Conference on Information Systems, AMCIS 2007: Reaching New Heights).
@inproceedings{efa95e85eb20409496ef0aec42927b9e,
title = "Using neural networks to forecast box office success",
abstract = "Predicting box office receipts of a particular movie has intrigued many researchers, domain experts and industry leaders as a challenging problem. In this paper, we report on the current status of a prediction system being built at the Institute for Research in Information Systems (IRIS) at Oklahoma State University since 1998. In our model, the forecasting problem is converted into a classification problem, that is, rather than forecasting the pinpoint estimate of box office receipts, a movie is classified into one of nine financial success categories, ranging from a {"}flop{"} to a {"}blockbuster.{"} The prediction results of different datasets representing different time windows and different combination of predictors are presented using average percent hit rate of bingo and oneaway predictions. In the latest tests the prediction results of artificial neural networks improved to almost 50{\%} on {"}bingo{"} and close to 90{\%} on {"}one-away{"}.",
keywords = "Box-office receipts, Classification, Data mining, Forecasting, Motion pictures, Neural networks, Performance measures, Prediction",
author = "Mike Henry and Ramesh Sharda and Dursun Delen",
year = "2007",
month = "12",
day = "1",
language = "English",
isbn = "9781604233810",
series = "Association for Information Systems - 13th Americas Conference on Information Systems, AMCIS 2007: Reaching New Heights",
pages = "1589--1597",
booktitle = "Association for Information Systems - 13th Americas Conference on Information Systems, AMCIS 2007",

}

Henry, M, Sharda, R & Delen, D 2007, Using neural networks to forecast box office success. in Association for Information Systems - 13th Americas Conference on Information Systems, AMCIS 2007: Reaching New Heights. Association for Information Systems - 13th Americas Conference on Information Systems, AMCIS 2007: Reaching New Heights, vol. 3, pp. 1589-1597, 13th Americas Conference on Information Systems, AMCIS 2007, Keystone, CO, United States, 10/08/07.

Using neural networks to forecast box office success. / Henry, Mike; Sharda, Ramesh; Delen, Dursun.

Association for Information Systems - 13th Americas Conference on Information Systems, AMCIS 2007: Reaching New Heights. 2007. p. 1589-1597 (Association for Information Systems - 13th Americas Conference on Information Systems, AMCIS 2007: Reaching New Heights; Vol. 3).

Research output: Chapter in Book/Report/Conference proceedingConference contribution

TY - GEN

T1 - Using neural networks to forecast box office success

AU - Henry, Mike

AU - Sharda, Ramesh

AU - Delen, Dursun

PY - 2007/12/1

Y1 - 2007/12/1

N2 - Predicting box office receipts of a particular movie has intrigued many researchers, domain experts and industry leaders as a challenging problem. In this paper, we report on the current status of a prediction system being built at the Institute for Research in Information Systems (IRIS) at Oklahoma State University since 1998. In our model, the forecasting problem is converted into a classification problem, that is, rather than forecasting the pinpoint estimate of box office receipts, a movie is classified into one of nine financial success categories, ranging from a "flop" to a "blockbuster." The prediction results of different datasets representing different time windows and different combination of predictors are presented using average percent hit rate of bingo and oneaway predictions. In the latest tests the prediction results of artificial neural networks improved to almost 50% on "bingo" and close to 90% on "one-away".

AB - Predicting box office receipts of a particular movie has intrigued many researchers, domain experts and industry leaders as a challenging problem. In this paper, we report on the current status of a prediction system being built at the Institute for Research in Information Systems (IRIS) at Oklahoma State University since 1998. In our model, the forecasting problem is converted into a classification problem, that is, rather than forecasting the pinpoint estimate of box office receipts, a movie is classified into one of nine financial success categories, ranging from a "flop" to a "blockbuster." The prediction results of different datasets representing different time windows and different combination of predictors are presented using average percent hit rate of bingo and oneaway predictions. In the latest tests the prediction results of artificial neural networks improved to almost 50% on "bingo" and close to 90% on "one-away".

KW - Box-office receipts

KW - Classification

KW - Data mining

KW - Forecasting

KW - Motion pictures

KW - Neural networks

KW - Performance measures

KW - Prediction

UR - http://www.scopus.com/inward/record.url?scp=84870164705&partnerID=8YFLogxK

M3 - Conference contribution

AN - SCOPUS:84870164705

SN - 9781604233810

T3 - Association for Information Systems - 13th Americas Conference on Information Systems, AMCIS 2007: Reaching New Heights

SP - 1589

EP - 1597

BT - Association for Information Systems - 13th Americas Conference on Information Systems, AMCIS 2007

ER -

Henry M, Sharda R, Delen D. Using neural networks to forecast box office success. In Association for Information Systems - 13th Americas Conference on Information Systems, AMCIS 2007: Reaching New Heights. 2007. p. 1589-1597. (Association for Information Systems - 13th Americas Conference on Information Systems, AMCIS 2007: Reaching New Heights).