Latent Semantic Analysis


Latent Semantic Analysis is used to define the valance attached to a given set of words where valance itself is defined as either being positive or negative on a scale from -5 to 5.

In this specific case, AFINN has been taken into use to analyze the development of each character throughout every episode of Star Wars. To do so, it is needed to analyze what each character has said throughout every episode based on the valance given from AFINN. Hereby, the manuscripts of each episode of Star Wars have been taken into use as these state which character have said what.

After having created a map of who said what and in which episode, it is possible to calculate the average valance for the given person for each episode by using this script (provided by Finn Årup Nielsen).

For instance, it is possible to see from the graph below that Anakin Skywalker (the red line) increasingly becomes more evil as he turns to the dark side. In the end of Episode 3 Anakin emerges as Darth Vader (the yellow line), who by the end of Episode 3, sees the flaws of his actions and turns away from the dark side.

You can change the graph yourself by either using the tag cloud or simply typing in different characters in the search field.

Tag cloud