Information Technology Reference
In-Depth Information
and [8]. Those oral history data were collected from three research participants
who had involved in the movie industry in the area from the 1910s to the 1930s.
Each research participant did not refer to each other in the data, but their
experiences are considered as connected in some cases. Therefore, extracting
personal cultures from all three research participants can give us some ideas
about collective culture of that area. The step-by-step procedures to use the
method are as follows:
1. Read the oral histories and collect direct spatiotemporal information from
them.
2. Enter the direct spatiotemporal information to the user dictionary.
3. Enter the indirect spatiotemporal information such as names of persons,
buildings, and events to the phrase-pattern dictionary.
4. Extract fragmented sentences from both morphological and syntactical anal-
yses by using the dictionaries.
5. Add the domain-topic repository to map the spatiotemporal information to
computer-readable data.
6. Add timestamp and geocode to the fragmented sentences.
7. Store the data to KC, and see how it works.
4.2 Analyzing Personal Culture
For morphological analysis, we add the user dictionary to the MeCab program
[4]. Furthermore, to make the syntactical analysis easier, we also add the phrase-
patterns to the program. After doing so, we prepare a tool for these two analyses
to split text into lines which can go through the line-by-line analysis. As an
experiment, we set the threshold parameter as three lines, which are treated
as a unit of fragmented sentences. With the domain-topic repository, we add
timestamp and geocode data to the fragmented sentences, extracted from the
first two analyses.
4.3 Visualizing and Exploring Collective Culture
After the procedures above, we make instances of the StoryFragment class from
the fragmented sentences with spatiotemporal information. The instances of Sto-
ryline class are generated based on the contextual information of the fragmented
sentences. Fig. 6 shows a result of plotting those instances on KC.
We could get the narrative data from three research participants in this test
case. Since some stories in their narratives were referring to a similar time and
place, KC's OPP detector could automatically find the intersection point of
their storylines in the virtual 3D space (see Fig. 7). With the assist of the OPP
detector, we could know a focal point to start exploring the sharable experiences
in the community, and get a clue to understand the indirect collective culture
regarding movie industries in Kyoto.
In this test case, KC found one OPP where three storylines from three different
research participants went through a similar time and place in the virtual 3D
