Movies are one of the most popular entertainments in our lives today. Some mathematicians found an interesting method to reveal all scenes during a movie, which is called “character-net”. Character-net was a graph that showed a social network by the different dialogs between characters. We named it G=(V, E, W). Each character is matched to a node( V), each dialog’s direction is matched to an edge(E), and each dialog’s time is matched to a weight(W).
The above figure is a simple case to represent a character-net, which only consists one dialog from character c1 to character c2. We called the character who starts the dialog “speaker”, and called the other one who receives the dialog “listener”. The arrow between c1 and c2 is the direction of the dialog. Also, we name w12 as the time that this dialog spends, which is the weight in the graph. The whole graph can describe one scene in a movie. Let’s imagine, if we need to construct all scenes through a movie, how many graphs should we use?
From the above figure, we can get some useful information. In this related movie, there are totally seven main characters. We classify a node with the most adjacent nodes and dialog directions as a major character. Oppositely, those who have the fewest adjacent nodes and dialog directions would be classified as extras. These information is very useful to measure each character’s importance in the movie. Also, they help us understand the specific character relation.
Usually, we divide movie roles into three categories: major role, minor role and extra. But how to distinguish them?
Here we bring in a new concept called “DC value”. It can be calculate as follows:
As we can see, characters below the average DC value are classified as extras. Characters over the average DC value are classified as major and minor roles. Then, the value that divides major and minor roles is the biggest gap among characters.
I think character-net makes a nice connection between art and science. On the one hand, it shapes an artistic mathematic model to make these programmed science theory more attractive. On the other hand, people can learn more scientific knowledge during their daily entertainments. If some of you love watching movies, try to draw a character-net graph to the related movie, it’s really interesting!
Source:
https://www.researchgate.net/publication/225765648_Social_network_analysis_in_a_movie_using_character-net