利用網路使用者評論預測電影票房之研究

電影已經成為民眾主要的休閒活動之一，票房紀錄也不斷創新高，但是電影的製作成本很高，若沒有詳盡的評估，便會造成虧損，而若能提前預測電影票房，便能讓製片公司、發行商和投資者更容易地做決策。電影賣座的原因有很多，導演、影星和宣傳等，而在資訊發達的現在，網路已經成為電影宣傳的主要管道之一，網路上的評論也會被民眾所看到，成為是否觀看該部電影的原因之一。本研究從IMDb蒐集2009年到2014年間，五種不同電影類型，總共1,658部電影的資料和第一週電影評論，使用SentiStrength和Stanford CoreNLP兩種情感分析工具將評論量化，並結合電影基本資料、外在環境因素等因素預測電影票房。使用10折驗證評估，並應用資料探勘軟體Weka中的M5P、線性迴歸和SMOreg三種預測技術建立電影票房預測模型，最後將其結果和沒有使用評論的電影票房預測結果進行比較，評估的指標包括相關係數、平均絕對誤差和均方根誤差。此外，本研究也找出對電影票房影響較大的因素。結果顯示，加入評論的預測結果會比沒有使用評論的預測結果來的更好，但是若將電影再依年份分開預測，會因為訓練資料集太小造成結果不理想。加入評論預測電影票房可以提升預測的準確率，評論中對票房影響最大的為評論的數量，評論內容則只會對特定類型的電影產生影響。

關鍵字

電影評論；情感分析；票房預測

並列摘要

Seeing movies has become the most popular leisure activity now, the box-office continue to break records, but the budget of movies are so high, if there were no evaluation, it may be huge loss. If we can predict box-office before it is released, film studios, film distribution and investors can make their decision more easily. There are many factors to influence box-office, directors, actors/actresses, etc. Now, Internet has become the most important way to transfer information, online reviews may be seen by users and take effect on them. Our research collects 2009 to 2014 movie data and reviews from IMDb, includes 5 genres and total 1,658 movies, we use SentiStrength and Stanford CoreNLP to analyse reviews, combined with movies data and external factor to predict box-office. We use M5P, Linear Regression and SMOreg methods to build prediction models, and compare the result without reviews and result with reviews. The result shows that the prediction models with reviews are better than without reviews, but if we spilt the data into different year, the results would be lower because the lack of data. Reviews can help the result of prediction, and the principal factor is the number of reviews, reviews can only influence specific genres of movies.

並列關鍵字

movie reviews ； sentiment analysis ； box-office prediction

參考文獻

Ba, S., & Pavlou, P. A. (2002). Evidence of the effect of trust building technology in electronic markets: Price premiums and buyer behavior. Management Information Systems Quarterly, 243-268.

Bai, X. (2011). Predicting consumer sentiments from online text. Decision Support Systems, 50(4), 732-742.

Basuroy, S., Chatterjee, S., & Ravid, S. A. (2003). How critical are critical reviews? The box office effects of film critics, star power, and budgets. Journal of Marketing, 67(4), 103-117.

Cantallops, A. S., & Salvi, F. (2014). New consumer behavior: A review of research on eWOM and hotels. International Journal of Hospitality Management, 36, 41-51.

Cao, Q., Duan, W., & Gan, Q. (2011). Exploring determinants of voting for the “helpfulness” of online user reviews: A text mining approach. Decision Support Systems, 50(2), 511-521.

國際替代計量

利用網路使用者評論預測電影票房之研究

未授權

主題瀏覽