電影總票房與日票房的預測模型對影院佈局以及前期投資優化具有重要作用。同時，隨著社群網路近年來快速的發展，社群聲量對電影票房的影響也被證明起到正相關作用。然而，影響日常票房的因素眾多，且票房時間序列的生命週期較短。針對上述兩個問題對模型構建帶來的挑戰，本文嘗試透過分別建立隨機森林的回歸模型以及長短期記憶網路模型（Long Short-Term Memory, LSTM）予以解決。通過建立隨機森林的回歸模型本文發現並提取了隱藏在眾多電影相關因數中的先決因子，如電影上映前即需要先行決定的“上映院數”。經實驗驗證，先決因子的發現有效提高了模型的準確度，也為後續的LSTM模型打下基礎。最後，本文通過單向LSTM模型，使用電視廣告投放資料與社群聲量等資料，對日票房序列進行預測，並著重比較投入廣告投放資料對LSTM模型的影響。經實驗驗證，未投入廣告資料的模型預測誤差為23.24%，而擁有廣告資料的模型預測誤差可增升為21.06%。

The prediction model of movie total box office and daily box office plays an important role on cinema allocation and the optimization of early investment. With the rapid development of social network, the influence of social listening on box office has been proved to play a critical role. The prediction of box office faces challenge because many factors affecting the box office, and the life cycle of box office sequence is usually short. Aiming at the challenges mentioned above, this paper proposed a prediction model by using the regression model of random forest and the Long Short-Term Memory (LSTM) model. By establishing the regression model of random forest for total box office, the prerequisite factor such as "the number of cinema" which is usually pre-determined before the movie comes out was extracted. Through experimental verification, the discovery of the prerequisite factor effectively improves the accuracy of the model and lays a foundation for the subsequent LSTM model. The one-way LSTM model was used to predict the daily box office by using data such as TV advertising and social listening indicator. The experimental results show that the prediction error of the model with no advertising data is 23.24%, and the error of the model with advertising data can be reduced to 21.06%.

參考文獻
[1] B. Egeli, "Stock market prediction using artificial neural networks," Decision Support Systems, vol. 22, pp. 171-185, 2003.
[2] W. D. Walls and J. McKenzie, "The changing role of Hollywood in the global movie market," Journal of Media Economics, vol. 25, no. 4, pp. 198-219, 2012.
[3] P. Nagamma, H. Pruthvi, K. Nisha, and N. Shwetha, "An improved sentiment analysis of online movie reviews based on clustering for box-office prediction," in International Conference on Computing, Communication & Automation, 2015, pp. 933-937: IEEE.
[4] Y. Ru, B. Li, J. Chai, and J. Liu, "A Daily Box Office Prediction Model with LSTM," in 2019 International Conference on Artificial Intelligence and Advanced Manufacturing (AIAM), 2019, pp. 548-551: IEEE.
[5] B. R. Litman, "Predicting financial success of motion pictures," The motion picture mega-industry, 1998.
[6] B. Liu, Web data mining: exploring hyperlinks, contents, and usage data. Springer Science & Business Media, 2007.
[7] A. Baloglu and M. S. Aktas, "BlogMiner: Web Blog Mining Application for Classification of Movie Reviews," presented at the 2010 Fifth International Conference on Internet and Web Applications and Services, 2010.
[8] S. B. Barnett, H. M. White, and M. Cerf, "Keep it simple stimuli: brain-vetted elements of movie trailers predict opening weekend ticket sales," ACR North American Advances, 2016.
[9] R. Faughnder, "Movie trailer makers multiply as online viewing of previews soars," Los Angeles Times, vol. 21, 2015.
[10] J. Eliashberg, A. Elberse, and M. A. Leenders, "The motion picture industry: Critical issues in practice, current research, and new research directions," Marketing science, vol. 25, no. 6, pp. 638-661, 2006.
[11] S. M. Choi, W.-N. Lee, and H.-J. Kim, "Lessons from the rich and famous: A cross-cultural comparison of celebrity endorsement in advertising," Journal of advertising, vol. 34, no. 2, pp. 85-98, 2005.
[12] S. Qiu and H.-Q. He, "Machine Learning- and Evidence Theory-Based Fraud Risk Assessment of China’s Box Office," IEEE Access, vol. 6, pp. 75619-75628, 2018.
[13] N. Quader, M. O. Gani, and D. Chaki, "Performance evaluation of seven machine learning classification techniques for movie box office success prediction," in 2017 3rd International Conference on Electrical Information and Communication Technology (EICT), 2017, pp. 1-6: IEEE.
[14] R. Dhir and A. Raj, "Movie Success Prediction using Machine Learning Algorithms and their Comparison," in 2018 First International Conference on Secure Cyber Computing and Communication (ICSCCC), 2018, pp. 385-390: IEEE.
[15] W. Zhang and S. Skiena, "Improving movie gross prediction through news analysis," in 2009 IEEE/WIC/ACM International Joint Conference on Web Intelligence and Intelligent Agent Technology, 2009, vol. 1, pp. 301-304: IEEE.
[16] A. Oghina, M. Breuss, M. Tsagkias, and M. De Rijke, "Predicting imdb movie ratings using social media," in European Conference on Information Retrieval, 2012, pp. 503-507: Springer.
[17] M. Ghiassi, J. Skinner, and D. Zimbra, "Twitter brand sentiment analysis: A hybrid system using n-gram analysis and dynamic artificial neural network," Expert Systems with applications, vol. 40, no. 16, pp. 6266-6282, 2013.
[18] S. Asur and B. A. Huberman, "Predicting the future with social media," in 2010 IEEE/WIC/ACM international conference on web intelligence and intelligent agent technology, 2010, vol. 1, pp. 492-499: IEEE.
[19] M. Cerrada, G. Zurita, D. Cabrera, R.-V. Sánchez, M. Artés, and C. Li, "Fault diagnosis in spur gears based on genetic algorithm and random forest," Mechanical Systems and Signal Processing, vol. 70-71, pp. 87-103, 2016.
[20] L. Breiman, "Random Forests," Machine Learning, journal article vol. 45, no. 1, pp. 5-32, Oct 01 2001.
[21] F. Eyben, F. Weninger, S. Squartini, and B. Schuller, "Real-life voice activity detection with lstm recurrent neural networks and an application to hollywood movies," in 2013 IEEE International Conference on Acoustics, Speech and Signal Processing, 2013, pp. 483-487: IEEE.
[22] L. Rahman, N. Mohammed, and A. K. Al Azad, "A new LSTM model by introducing biological cell state," in 2016 3rd International Conference on Electrical Engineering and Information Communication Technology (ICEEICT), 2016, pp. 1-6: IEEE.
[23] A. Ainslie, X. Drèze, and F. Zufryden, "Modeling movie life cycles and market share," Marketing science, vol. 24, no. 3, pp. 508-517, 2005.
[24] Y. Liu, "Word of mouth for movies: Its dynamics and impact on box office revenue," Journal of marketing, vol. 70, no. 3, pp. 74-89, 2006.
[25] L. Wang and J.-m. Jia, "Forecasting box office performance based on online search: Evidence from Chinese movie industry," Systems Engineering-Theory Practice, vol. 34, no. 12, 2014.
[26] S. Hochreiter and J. Schmidhuber, "LSTM can solve hard long time lag problems," in Advances in neural information processing systems, 1997, pp. 473-479.
[27] K. Kawakami, "Supervised sequence labelling with recurrent neural networks," Ph. D. dissertation, vol. 66, 2008.
[28] H. Salehinejad, S. Sankar, J. Barfett, E. Colak, and S. Valaee, "Recent advances in recurrent neural networks," arXiv preprint arXiv:.01078, 2017.
[29] J. C. B. Gamboa, "Deep learning for time-series analysis," Machine Learning, vol. 152, 2017.