CEU Electronic Theses and Dissertations, 2019
| Author | Zinoveva, Aleksandra |
|---|---|
| Title | Can Box Office Success Be Predicted Before a Movie's Release? |
| Summary | Due to the motion picture industry being a very risky one with high sunk costs, film studios executives around the world are trying to find better ways to predict whether their films are going to be successful even as early as in the initial stages of development. As only the big players, such as Hollywood studios, have unobstructed access to the real market data, in this thesis I investigate the possibility of predicting a movie’s success based on publicly available data on previously released films, which is usually the only information an independent film producer is able to obtain. For this, I fit a Logistic Regression model to the data for 4,727 movies released in 1982-2016. The results show that predicting algorithm does not perform very well, correctly classifying only 62% of films when working with this limited information; however, such factors as movie’s budget, facts of belonging to a franchise or being produced with participation of an American company, have proven themselves to be important for distinguishing between potentially successful and unsuccessful films. |
| Supervisor | Muço, Arieda |
| Department | Economics MA |
| Full text | https://www.etd.ceu.edu/2019/zinoveva_aleksandra.pdf |
Visit the CEU Library.
© 2007-2021, Central European University