Information Technology Reference
In-Depth Information
culture is externalized. In other words, shared within the community, the per-
sonal culture affects the collective counterpart. On the other hand, collective
culture is internalized to the individual's sense and personal culture.
Since both personal and collective cultures have intangible, complex, and am-
biguous characteristics, capturing them in socio-cultural contexts is one of chal-
lenging tasks for researchers. To understand this contextual feature of culture,
we specifically focus on spatiotemporal information as a key to capture personal
and collective cultures. When people recount their personal experience, they
frequently give utterances directly or indirectly related to spatiotemporal infor-
mation. We also found different people sometimes refer to the same location or
time in their stories. Intersection of such stories could be considered as a clue to
understand collective culture. Although personal culture may not be external-
ized directly, we could have a glimpse of the collective culture from collection of
independent stories as a result of indirect externalization of personal culture.
In the following, we propose a method to extract the information for personal
culture from transcribed oral histories. The extraction procedure is followed by
the introduction of the KACHINA CUBE (KC) system, a web-based platform
for visualizing and mining regional collective culture. We, then, test KC with a
case of collective culture from three research participants, regarding the once-
flourishing film industry in Kyoto. Finally, we evaluate our methods and the
platform by comparing to the related works.
2 Mining Personal and Collective Cultures
This section elucidates the procedure of extracting information of personal and
collective culture from the narrative data. In reality, the extraction process fol-
lows the research that involves collecting narratives from research participants. It
should be also mentioned that the linguistic processing in this section is partially
assuming the Japanese textual data transcribed from oral-history interviews with
the regional community members in a certain area of Kyoto, although the main
part of the following procedure would be applicable to other languages.
2.1 Extracting Personal Culture
We consider the spatiotemporal information appeared in narrative data would
be a guide to split them into chunks related to personal culture. Fig.1 explains
procedure to extract personal culture from the narrative data. The procedure
contains morphological analysis, syntactical analysis, and spatiotemporal anal-
ysis. After obtaining the narrative data, researchers first perform morphological
analysis on them, splitting them into morpheme units. Using both the basic and
user dictionaries, we can detect morpheme units from the original transcripts. To
explain this in more detail, the user dictionary covers miner or specific terms that
are not registered in the basic dictionary. For example, “Uzumasa-Eigamura,”
or Uzumasa Movie Village, indicates one specific location, which the basic dic-
tionary might not register as a proper noun. Such names of locations, events,
persons, and buildings are considered as entries for the user dictionary.
