Supporting Exploratory Information Access Based on Comic Content Information

Abstract

SummaryThe purpose of this research is supporting information access based on the contents of comic books. To meet this purpose, it is necessary to obtain information related to the story and the characters of a comic. We propose a method to extract information from reviews on the Web by using term frequency-inversed document frequency (TF-IDF) method and hierarchical Latent Dirichlet Allocation (hLDA) method, which intends to solve the problem. By using these methods, we build a prototype system for exploratory comic search. We conducted a user study to observe how a participant use the system. The user study showed that the system successfully supported the participants to find interesting unread comics.

Designing a Comic Exploration System Using a Hierarchical Topic Classification of Reviews

“…As mentioned above, in this research, we concluded that the named entity in the review sentence is information that can not be referred to for grasping the contents, and such part of speech is excluded from the analysis target in advance. However, it was confirmed that some named entities were classified as general nouns in the analysis results [9]. This is because the MeCab 5 morphological analysier was unable to analyze all the existing named entities during the comic review.…”

“…In our previous research, we employed "preferable title" as the input query to reduce the cost of query generation [9]. As mentioned in the previous section, query patterns that respond to content-related requests may vary.…”

“…In the previous research [9], the words included in the topic were generated using the characteristic word of each comic extracted using the TF-IDF method and the topic generated using the hLDA method. When the user searches for a comic using the comic search system that contained words that were not recognized by the system, it was a burden on the user to guess the contents of the comic (see Figure 1).…”

“…Reviews of comics often contain both objective information (e.g., abstract of the story, episodes, and story settings) and subjective information (e.g., likes/dislikes and construals derived from the story). In a previous study, we proposed a method to extract content information from a comic indirectly by using the comic's reviews [6,9]. The proposed method evaluated comics more precisely than the conventional search services that only take into account genre information.…”

“…On the other hand, among the feature words extracted by the previous method [9] included many expressions based on unknown words defined in the comic, the characters that appeared in the comic, and the story of the comic are included. In this paper, these expressions are called named entities.…”