Fig 1.
Network Extracted from the book “Harry Potter and the Philosopher’s Stone” [1] with sentiment scores (Produced with NetworkX [50]).
Fig 2.
Schematic representation of one interaction between characters.
Fig 3.
Text extract of a conversation from “Harry Potter and The Philosopher’s Stone” [1] displaying the different elements manipulated (i.e. Narration, dialogs, context, conversation,…) and their relation with each other.
Fig 4.
Schematic view of the entire algorithm’s process.
Fig 5.
Dialog spacing distribution in “Harry Potter and the Philosopher’s Stone” [1].
Fig 6.
Dialog spacing distribution in “A Game of Thrones” [8].
Fig 7.
Dialog spacing distribution in “Les Misérables” [47].
Fig 8.
Dialog spacing distribution in “Harry Potter and the Philosopher’s Stone” [1] with the computed threshold.
Fig 9.
Example of input sentence and it’s related metadata.
Table 1.
Part-of-speech tags.
Fig 10.
Example of the first step of the speaker’s identification process.
Locating the subject of the sentence.
Table 2.
Rate of speaker identification (SIR).
Fig 11.
Differential between the extracted and correct networks in “Harry Potter and the Philosopher’s Stone” [1].
Red edges are incorrect edges, green edges were identified correctly.
Fig 12.
Context network #64 in “Harry Potter and the Philosopher’s Stone” [1].
Fig 13.
Example of incremental networks for “Harry Potter and the Philosopher’s Stone” [1].
On the left, context 0 to 498. On the right, context 0 to 1670.
Fig 14.
Network Extracted from the book “Harry Potter and the Order of the Phoenix” [5].
Fig 15.
Degree distribution for all Harry Potter books.
Table 3.
Exponent of the power approximation.
Fig 16.
Degree distribution in the full “Harry Potter” series plus the power approximation.
Fig 17.
Clustering coefficient in the Harry Potter books.
Fig 18.
Average clustering coefficient for the seven “Harry Potter” books processed as a single entity.
Table 4.
Average clustering coefficient for each book & clustering coefficient for a similar random network.
Fig 19.
Preferential attachment in “Harry Potter and The Philosopher’s Stone” [1].
This plot represents the probability, given the degree of a given node at a given time, that a new node being added to the social network will have a connection to that node.
Fig 20.
Preferential attachment in “Harry Potter and The Chamber of Secrets” [2].
Fig 21.
Dendrogram for the hierarchical clustering of the books using the Ward variance minimization algorithm.
Table 5.
Clusters generated with K-Means, K = 10.