Performing arts and movies have become commercial products with high profit and great market potential.
Previous research works have developed comprehensive models to forecast the demand for movies. However,
they did not pay enough attention to the decision support for performing arts which is a special category unlike
movies. For performing arts with high-dimensional categorical attributes and limit samples, determining ticket
prices in different levels is still a challenge job faced by the producers and distributors. In terms of these difficulties,
factorization machine (FM), which can handle huge sparse categorical attributes, is used in this work
first. Adaptive stochastic gradient descent (ASGD) and Markov chain Monte Carlo (MCMC) are both explored
to estimate the model parameters of FM. FM with ASGD (FM-ASGD) and FM with MCMC (FM-MCMC) both
can achieve a better prediction accuracy, compared with a traditional algorithm. In addition, the multi-output
model is proposed to determine the price in multiple price levels simultaneously, which avoids the trouble of the
models’ repeating training. The results also confirm the prediction accuracy of the multi-output model, compared
with those from the general single-output model.
XU Yong1 (徐勇), TANG Qian2 (唐倩), HOU Linzao2 (候林早), LI Mian2* (李冕)
. Decision Model for Market of Performing Arts with Factorization Machine[J]. Journal of Shanghai Jiaotong University(Science), 2018
, 23(1)
: 74
-84
.
DOI: 10.1007/s12204-018-1912-2
[1] SHANKLIN W. What businesses can learn from themovies [J]. Business Horizons, 2002, 45(1): 23-28.
[2] ELBERSE A, ELIASHBERG J. Demand and supplydynamics for sequentially released products ininternational markets: The case of motion pictures[J].Marketing Science, 2003, 22(3): 329-354.
[3] SQUIRE J E. The movie business book [M]. New York:Simon and Schuster, 2004.
[4] KIM T, HONG J, KANG P. Box office forecasting usingmachine learning algorithms based on SNS data[J]. International Journal of Forecasting, 2015, 31(2):364-390.
[5] DELEN D, SHARDA R, KUMAR P. Movie forecastGuru: A Web-based DSS for Hollywood managers [J].Decision Support Systems, 2007, 43(4): 1151-1170.
[6] AINSLIE A, DR`EZE X, ZUFRYDEN F. Modelingmovie life cycles and market share [J]. Marketing Science,2005, 24(3): 508-517.
[7] LEE K J, CHANGW. Bayesian belief network for boxofficeperformance: A case study on Korean movies [J].Expert Systems with Applications, 2009, 36(1): 280-291.
[8] ELIASHBERG J, SHUGAN S M. Film Critics: Influencersor predictors? [J]. Journal of Marketing, 1997,61(2): 68-78.
[9] CHINTAGUNTA P K, GOPINATH S,VENKATARAMAN S. The effects of online userreviews on movie box office performance: Accountingfor sequential rollout and aggregation across localmarkets [J]. Marketing Science, 2010, 29(5): 944-957.
[10] DUAN J, DING X, LIU T. A Gaussian copula regressionmodel for movie box-office revenues prediction [J].Science China, 2017, 60(9): 092103.
[11] LEE Y, KIM S H, CHA K C. A generalized Bassmodel for predicting the sales patterns of motion pictureshaving seasonality and herd behavior [J]. Journalof Global Scholars of Marketing Science, 2012, 22(4):310-326.
[12] ZHOU Y, ZHANG L, YI Z. Predicting movie box-officerevenues using deep neural networks [J]. Neural Computing& Applications, 2017(3): 1-11.
[13] ZHANG L, LUO J, YANG S. Forecasting box officerevenue of movies with BP neural network [J]. ExpertSystems with Applications, 2009, 36(3): 6580-6587.
[14] BAUMOL W J, BOWEN W G. Performing arts, theeconomic dilemma: A study of problems common totheater, opera, music and dance [M]. New York: GreggRevivals, 1993.
[15] RENDLE S. Factorization machines [C]//IEEE InternationalConference on Data Mining. [s.l.]: IEEE,2010: 995-1000.
[16] RENDLE S. Factorization machines with libFM [J].ACM Transactions on Intelligent Systems and Technology,2012, 3(3): 57.
[17] RUDER S. An overview of gradient descent optimizationalgorithms [EB/OL]. (2017-09-19). [2016-01-09].http://sebastianruder.com/optimizing-gradient-descent/index.html.
[18] RENDLE S. Learning recommender systems withadaptive regularization [C]//ACM International Conferenceon Web Search & Data Mining. Seattle, Washington:ACM, 2012: 133-142.
[19] ANDRIEU C, FREITAS N D, DOUCET A, et al. AnIntroduction to MCMC for machine learning [J]. MachineLearning, 2003, 50(1/2): 5-43.
[20] BESAG J. An introduction to Markov chain MonteCarlo methods [C]//Mathematical Foundations ofSpeech and Language Processing. New York: Springer,2004: 247-270.
[21] RENDLE S. Social network and click-through predictionwith factorization machines [EB/OL]. (2017-09-19). http://kdd2012.sigkdd.org.
[22] BORCHANI H, VARANDO G, BIELZA C, et al. Asurvey on multi-output regression [J]. WIREs DataMining and Knowledge Discovery, 2015, 5(5): 216-233.
[23] ZHANG W, LIU X, DING Y, et al. Multi-output LSSVRmachine in extended feature space [C]//IEEEInternational Conference on Computational Intelligencefor Measurement Systems and Applications.[s.l.]: IEEE, 2012: 130-134.
[24] BREIMAN L. Random forests [J]. Machine Learning,2001, 45(1): 5-32.
