=Paper=
{{Paper
|id=Vol-2865/poster5
|storemode=property
|title=Snippets of Folk Legends: Adapting a Text Mining Tool to a Collection of Folk Legends
|pdfUrl=https://ceur-ws.org/Vol-2865/poster5.pdf
|volume=Vol-2865
|authors=Maria Skeppstedt,Rickard Domeij,Fredrik Skott
|dblpUrl=https://dblp.org/rec/conf/dhn/SkeppstedtDS20a
}}
==Snippets of Folk Legends: Adapting a Text Mining Tool to a Collection of Folk Legends==
https://ceur-ws.org/Vol-2865/poster5.pdf
Please contact the conference organizers at sanita.reinsone@lulfmi.lv if you require an accessible file, as the files provided by ConfTool Pro to
reviewers are filtered to remove author information, and this filter technique unfortunately cannot preserve screen reader information. However,
the filter can easily be disabled by the organizers.
Snippets of Folk Legends:
Adapting a Text Mining Tool
to a Collection of Folk Legends?
Maria Skeppstedt, Rickard Domeij, and Fredrik Skott
The Institute for Language and Folklore, Sweden
firstname.lastname@isof.se
Abstract. A topic modelling tool was adapted to requirements for a collection
of Swedish folk legends. To offer an overview of a list of folk legend texts, which
had been automatically extracted by the topic modelling tool, snippet text ver-
sions of the folk legends were displayed. The snippets were constructed from the
full-text versions of the legends using the sentences most relevant to the topics
extracted by the topic modelling algorithm. In addition, collection-adapted data
was constructed for performing a pre-processing of the folk legend texts, before
they were submitted to the topic modelling algorithm. This data consisted of a
collection-adapted stop word list and word lists for improving the quality of clus-
ters of semantically similar words.
Keywords: Folk Legends, Topic Modelling, Text Analysis.
1 Introduction
Topic modelling has been shown to be a useful text mining technique for automatically
identifying recurring information in large text collections. The technique has been ap-
plied on text collections belonging to a wide range of genres, e.g. news paper text [3],
open-ended survey questions [2], Internet discussion forums [12], micro blogs [14],
student essays [5] and folk legends [6, 11].
The topic modelling algorithm automatically analyses a text collection in search for
topics that often recur in the text. There are a number of topic modelling visualisation
tools, for instance, visualisations that focus on relations between the automatically de-
tected topics and the terms that represent these topics [1, 4]. We here instead use, and
develop, a tool that builds on previous research which shows the potential in performing
a manual analysis of the texts in which the topics detected are present. This tool, called
Topics2Themes1 [13], therefore also visualises the extracted texts and their associations
and provides support for performing a manual analysis of these texts. The tool also pro-
vides a pre-processing step using a word2vec model [8], in which several words can
be organised into groups based on semantic similarity, and treated as one concept. This
is achieved through an automatic clustering of the word2vec vectors that represent the
words included in the text collection.
?
Supported by the Swedish Research Council (project number 2017-00626).
1
The code for the tool is available at: https://github.com/mariask2/topics2themes.
Copyright © 2021 for this paper by its authors. Use permitted under
Creative Commons License Attribution 4.0 International (CC BY 4.0).
243
Topics2Themes has previously been applied to posts in discussion forum threads
[13], to micro blogs [10], and to research article paragraphs [9]. We recently applied
the tool to a text collection consisting of around 10,000 Swedish folk legends [11]. The
tool had previously been applied on English and Japanese texts, and to apply it on a
Swedish text collection we needed to configure it to use a Swedish stop word list and
a Swedish word2vec model. Apart from using standard Swedish stop words, we also
added words typical to the text collection to the stop word list, e.g. words that use older
spelling variants, and old grammatical forms.
Texts extracted by the topic modelling algorithm as most closely associated with
the topics detected are displayed by Topics2Themes in a scrollable list. By scrolling
through this text list, the user is meant to be given an overview of the content extracted.
However, despite most of the folk legend texts being relatively short, many of them are
longer than the very short types of texts for which the tool was originally developed.
This has the effect that the overview, meant to be provided by the scrollable list, is lost.
To achieve the same overview in those cases, we therefore added the functionality to
– instead of displaying the full texts in the scrollable list – display a short snippet of
the text. In addition to this technical adaption of the tool, we also (i) further adapted
the stop word list to the folk legends collection by adding additional collection-specific
words, and (ii) performed a manual correction of the automatically constructed concept
clusters.
2 Topics2Themes Applied on the Folk Legend Collection
Topics2Themes was applied on a text collection containing around 10,000 records of
folklore from the archive at the Institute for Language and Folklore in Gothenburg. The
material mainly consists of folk legends, i.e. typically narratives about our perceptions
of the world around us, about events in the past, and about everyday life manifestations
of that, which we perceive as supernatural. The legends are usually relatively short
and monoepisodic, have a fixed structure, and have been transmitted verbally from one
generation to another.
Most of the records have been collected during the Interwar Period, through inter-
views with older people in the countryside in western Sweden. The records either con-
tain (manual or HTR-based) transcriptions of handwritten text, or text obtained from
printed material using OCR. Most of the material is written in Standard Swedish, al-
though occasional dialectal expressions, as well as words written with older spelling
variants, occur in the texts.
The Topics2Themes tool uses the topic modelling algorithm Non-negative matrix
factorisation [7], and then represents each topic detected by the algorithm through, (i) a
list of texts in which the topic is present, and (ii) a list of terms that frequently occur in
these texts. We re-ran the topic modelling algorithm 50 times on our collection of folk
legends, and only topics stable enough to be included in all re-runs were retained. This
resulted in that 19, out of 25 requested, topics were retrieved by the tool.
Figure 1 shows the graphical user interface of Topics2Themes, when applied on
the folk legend collection. The tool displays information in four different panels. The
Topics panel, i.e. the second panel from the left, lists the topics that have been automat-
244
ically extracted by the topic modelling algorithm. A topic element in the Topics panel is
represented by a list element that contains the three terms most closely associated with
the topic, e.g. “kyrka – bygga – stannade” (“church – build – stayed”). The Terms and
Texts panels list the terms and texts, respectively, that are associated with the extracted
topics. We configured the tool to retrieve 10 terms and 40 texts for each topic. That is,
for each topic detected, the Terms panel displays 10 terms, and the Texts panel displays
40 texts. Term-topic and topic-text associations are visualised by lines that connect the
list elements. Finally, to the Themes panel (only shown partly in the figure), the user can
add recurring themes that are identified when the extracted texts are manually analysed.
That is, while the first three panels stem from an automatic process, the content of the
Themes panel is to be manually added by a user.
The stop word list that was used for pre-processing the texts in the folk legends col-
lection has been iteratively expanded. This process consists of first using Topics2Themes
to construct a topic model, and thereafter of using the Terms panel to examine terms
extracted (as well as the Texts panel to examine their textual contexts), in order to de-
termine which terms are suitable as stop words. The process is then started over, using
a list expanded with the newly detected stop words when constructing the new model.
This iterative process has so far resulted in a stop word list consisting of 542 words.
For creating clusters of semantically similar words, we used a pre-trained word2vec
continuous skipgram model2 and the automatic clustering functionality of the Top-
ics2Themes tool. We then used the functionality provided by the tool for manually
correcting the automatically constructed clusters. This is carried out by providing the
tool with two types of word lists, (i) a list of words that are to be removed from the
automatic clusters (since the automatic process assigns them to a semantic cluster to
which they do not belong), and (ii) a list of manually constructed word clusters. Our
manually constructed clusters were often automatic clusters that we expanded with ad-
ditional, semantically similar words, or automatic clusters that we divided into smaller,
more semantically coherent groups of words. The construction of the word lists for im-
proving the clustering is also an iterative process. We have so far detected 571 words
to exclude from the automatic clustering, and constructed 118 manual clusters based
on the automatic ones. A total of 369 concept clusters were used when pre-processing
the texts before sending them to the topic modelling algorithm. Examples of concept
clusters are shown in the Terms panel. For instance, the first element consists of differ-
ent inflections of “church”, and the tenth element consists of a large cluster of words
referring to body parts. The advantage of using semantic clustering instead of methods
based on morphology – e.g. lemmatisation – is that semantically similar words can be
conflated into one concept, regardless of their lexical instantiations.3
2
The model has been trained on the Swedish CoNLL17 corpus – which contains 3,010,472
tokens – by the Language Technology Group at the University of Oslo.
The model is available at: http://vectors.nlpl.eu/repository/.
3
The topic model configuration, as well as the manually constructed word list and clusters is
available at: https://github.com/mariask2/swedish-folk-legends.
245
Fig. 1. Topics2Themes applied on a collection of Swedish folk legends. The topic “kyrka – bygga
– stannade” (“church – build – stayed”) is displayed with a blue background in the center panel,
which indicates that it has been selected by the user. The snippet versions of the texts most closely
associated with the selected topic are shown to the right, while associated terms are indicated with
a bold border to the left.
246
3 Producing the Snippet Text for the Extracted Sentences
The snippet text is produced by only retaining those sentences that contain terms that
the topic modelling algorithm has extracted as typical to the topics. A sentence – or a
sequence of sentences – that does not contain any of these terms is replaced by ellipsis,
as shown in Figure 1. The Topics2Themes configuration file lets the user specify the
maximum number of sentences to be included in the snippet text. This has the effect
that only the first sentences in the original text that contain a term will be added to the
text snippet, until the snippet has reached the maximum length specified. However, it is
also made sure that the snippet text includes at least one context sentence for each one
of the terms occurring in a text. Therefore, when more terms occur in the text than the
maximum snippet length specified, the length of the snippet will instead be determined
by the number of unique topic model-extracted terms occurring in the text.
The interface of Topics2Themes shows the snippet version of the text by default in
the Texts panel. There are two methods by which the user can make the tool display
the full-text version, (i) either by double-clicking on one specific list element to make
the tool show the full-text version for this specific text, or (ii) by clicking on the arrow
button in the header of the Texts panel to make the tool show the full-text versions for
all texts.
4 Future Work
We have now carried out the technical developments required for adapting the Top-
ics2Themes tool to our collection of folk legends. We have also provided the tool with
collection-adapted data in the form of an adapted stop word list and word lists for im-
proving the word2vec-based word clusters. As the next step, we will therefore use Top-
ics2Themes for carrying out a computer-assisted analysis of the folk legend collection.
That is, we will use the tool for identifying recurring themes in Swedish folk legends,
i.e. themes that might (i) reflect already established classification systems, as well as
(ii) provide new dimensions on how legends can be categorised in folkloristics.
References
1. Alexander, E., Kohlmann, J., Valenza, R., Witmore, M., Gleicher, M.: Serendip: Topic
model-driven visual exploration of text corpora. In: 2014 IEEE Conference on Visual Ana-
lytics Science and Technology (VAST). pp. 173–182 (2014)
2. Baumer, E.P.S., Mimno, D., Guha, S., Quan, E., Gay, G.K.: Comparing grounded theory and
topic modeling: Extreme divergence or unlikely convergence? Journal of the Association for
Information Science and Technology 68(6), 1397–1410 (2017)
3. Blei, D.M.: Topic modeling and digital humanities. Journal of Digital Humanities (2012)
4. Chuang, J., Manning, C.D., Heer, J.: Termite: Visualization techniques for assessing textual
topic models. In: Proceedings of the International Working Conference on Advanced Visual
Interfaces. pp. 74–77. ACM, New York, NY, USA (2012)
5. Ferrara, A., Montanelli, S., Petasis, G.: Unsupervised detection of argumentative units
though topic modeling techniques. In: Proceedings of the 4th Workshop on Argument Min-
ing. pp. 97–107. Association for Computational Linguistics, Copenhagen, Denmark (2017)
247
6. Karsdorp, F., den Bosch, A.V.: Identifying motifs in folktales using topic models. In: Pro-
ceedings of the 22 Annual Belgian-Dutch Conference on Machine Learning (2013)
7. Lee, D.D., Seung, H.S.: Algorithms for non-negative matrix factorization. In: Advances in
neural information processing systems. pp. 556 – 562 (2001)
8. Mikolov, T., Chen, K., Corrado, G., Dean, J.: Efficient estimation of word representations in
vector space (2013), http://arxiv.org/abs/1301.3781, cite arxiv:1301.3781
9. Skeppstedt, M., Ahltorp, M., Eriksson, G., Domeij, R.: Annotating risk factor mentions in
the COVID-19 Open Research Dataset. In: CLARIN2020 Book of Abstracts (2020)
10. Skeppstedt, M., Ahltorp, M., Kucher, K., Kerren, A., Rzepka, R., Araki, K.: Topic modelling
applied to a second language: A language adaptation and tool evaluation study. In: Selected
Papers from the CLARIN Annual Conference 2019. vol. 172:17, pp. 145–156. Linköping
Electronic Conference Proceedings (2020)
11. Skeppstedt, M., Domeij, R., Skott, F.: Adapting a topic modelling tool to the task of finding
recurring themes in folk legends. In: Proceedings of the Digital Humanities in the Nordic
Countries. pp. 388–392. CEUR Workshop Proceedings (2020)
12. Skeppstedt, M., Kerren, A., Stede, M.: Vaccine hesitancy in discussion forums: Computer-
assisted argument mining with topic models. In: Building Continents of Knowledge in
Oceans of Data: The Future of Co-Created eHealth. pp. 366–370. No. 247 in Studies in
Health Technology and Informatics, IOS Press (2018)
13. Skeppstedt, M., Kucher, K., Stede, M., Kerren, A.: Topics2Themes: Computer-Assisted Ar-
gument Extraction by Visual Analysis of Important Topics. In: Proceedings of the LREC
Workshop on Visualization as Added Value in the Development, Use and Evaluation of Lan-
guage Resources. pp. 9–16 (2018)
14. Surian, D., Nguyen, D.Q., Kennedy, G., Johnson, M., Coiera, E., Dunn, A.G.: Characterizing
Twitter discussions about HPV vaccines using topic modeling and community detection. J
Med Internet Res 18(8), e232 (2016)
Powered by TCPDF (www.tcpdf.org)