Association Meaning in Identifying Pleasant Tweets
Diyas Puspandari
1,2
, Syihabuddin Syihabuddin
1
and Wawan Gunawan
1
1
School of Postgraduate, Universitas Pendidikan Indonesia, Bandung, Indonesia
2
School of Computing, Telkom University, Bandung, Indonesia
Keywords: Association Meaning, Semantic Analysis, Pleasant Meaning.
Abstract: Social media, such as Twitter, can be used by the government to analyze netizens’ opinions. From
thousands of tweets, certainly it will take time if the analysis is conducted manually. Therefore, an analysis
tool is needed to know current opinion in a fast and real-time way. The purposes of this research are: 1) to
describe the association meaning formed from pleasant tweets in Bandung, 2) to build a classification
system based on semantic analysis to find out pleasant tweets of society in Bandung in near real-time way.
The step conducted was by filtering words from the tweets to examine which words experience changes in
meaning (have pleasant association meaning), mainly on words which convey places then change by getting
additional meaning because of undergoing association process. The results of this research are 1) the
database of association meaning of tweets from society in Bandung, 2) the application to analyze pleasant
tweets in near real-time way based on association meaning.
1 INTRODUCTION
Research on the use of Twitter for the measurement
of happiness index has been widely studied. Curini,
Iacus, and Canova (2014), built a measure of the
level of happiness at the provincial level using
statistical techniques which is innovative with data
from Twitter Italian society. Another study related to
happiness index has been done by Mitchell, et.al.
(2013) which the results suggest that social media
can potentially be used to estimate real-time rates
and changes in measuring the population scale.
Ponilan, Herdiani and Selviandro (2016) measured
the happiness index in Bandung on social media
Twitter using Top-Down Hierarchy ontology
method. The research described how to build
ontology to calculate happiness index of Bandung
citizens. However, there is no study that analyzes the
tweets semantically to know the happiness index.
Because the happiness index factor is so complex,
this research just analyzing the semantic meaning of
the tweets to identify the pleasant meaning. Based
on this gap, the purposes of this research are: 1) to
describe the association meaning formed from pleasant
tweets on tweets in Bandung, 2) to build a
classification system based on semantic analysis to
find out pleasant tweets of society in Bandung in
near real-time way.
Commonly, tweets contain statement and
frequently involve adverb of places, for example
“walking around@Lembang”. The word Lembang in
this case is not only to declare the name of region
instead it gets additional meaning such as exciting
place for nature tour and culinary tour. In
consequence, the data will be analysed based on
association meaning from the tweets.
Association meaning can occur in several types
of words, such as noun and adverb of place. In this
research, however, the analysis is limited to adverb
of place.
2 ASSOCIATION MEANING
Leech (in Chaer, 2012) divides meaning into
conceptual and associative. Conceptual meaning is
from lexeme which is without any correlation to its
context and other associations. Another division,
associative meaning is from lexeme which has
correlation with the thing out of the language itself.
Based on those definitions, association is the
relationship between the original meaning, meaning
in the environment where the original word grows
with the new meaning; i.e. meaning in the
environment in which the word is transferred into
language usage. Between the old meaning and the
446
Puspandari, D., Syihabuddin, S. and Gunawan, W.
Association Meaning in Identifying Pleasant Tweets.
DOI: 10.5220/0007168904460449
In Proceedings of the Tenth Conference on Applied Linguistics and the Second English Language Teaching and Technology Conference in collaboration with the First International Conference
on Language, Literature, Culture, and Education (CONAPLIN and ICOLLITE 2017) - Literacy, Culture, and Technology in Language Pedagogy and Use, pages 446-449
ISBN: 978-989-758-332-2
Copyright © 2018 by SCITEPRESS – Science and Technology Publications, Lda. All rights reserved

new meaning, there is a close relationship. The
association meaning can be linked to a place or
location. If we mention Senayan, people must know.
Its meaning does not refer to the place, but what is
there. Senayan is famous for its complex of sports
activities. That is why if people say, "Let’s go to
Senayan" then what is meant is not to Senayan area,
but it means to watch the game (related to sports) in
Senayan. Here it is the change of meaning seen, that
is not the place, but the things related to the place
(Pateda, 2010).
Associative meaning is similar to the
symbolization used by a certain language society to
state other concept which is related to moral value
and principles, therefore, it will be associated with
taste value (Chaer, 2002).
The shifting and changing of the meaning based
on the above explanation can effect on assessing the
word meaning. This assessment can be negative
(something dislike or peyorative) or positive
(something like or ameliorative). This arises based
on the use of word meaning in context and situation
(Parera, 2004).
This study focuses on the association meaning
which has assessment on pleasant meaning.
According to KBBI (2016), pleasant means 1)
making fun, 2) making joy, 3) exciting, 4) satisfying,
5) attracting (heart), and 6) being happy (satisfied
and so on), 7) like.
Associated with the association meaning, words
such as bioskop, Burangrang, and alun-alun are the
places with association meaning. This is owing to
the fact that the word bioskop does not only refer to
the building alone but also has additional meaning as
a pleasant place to watch movies, Burangrang also
does not only refer to the name of the road but it
means areas of fun for culinary tour, and alun-alun
is not only the land roomy, but it gets extra pleasant
meaning because it usually becomes a place of
sightseeing, culinary tour, and shopping.
3 METHODS
The tweets taken as data were simply the tweets
uploaded from Bandung city area. The data were
taken from one week, from Monday to Sunday. Data
collection was done in the hope that the tweets taken
could represent all days, both weekdays and
weekends, so that places with pleasant association
meaning were found all days. This one-week limit
was also due to Twitter's terms. Unfortunately, at the
time of the research, there was a disruption from the
Twitter so that the tweet data could be taken in only
1 day, i.e. September 26, 2017.
The data were taken from Bandung area from 1
geocode with Trans Studio Mall radius of 8 km
(about 200 km2). The tweets were filtered until it
generated tweets containing words with association
meaning. The next step was to rank words with
association meaning to know the public's opinion in
general whether those words have pleasant
association meaning or not. Then, a list of words that
had been ranked were analysed semantically, to find
out that the word means a pleasant association. The
last step was to build a system based on data from
the list of words with association meaning and to test
system accuracy.
Below is the flow diagram of the study portrayed
with the description in figure 1.
Figure 1: Flow diagram of the study.
The research method was done with the
following steps:
a. Literary Study
b. Data Collection
c. Data Selection (words filtering)
d. System Planning
At this stage, analysis and system planning was
conducted to classify the tweets into the category
of pleasant or not. The design includes
determining the required features and techniques
for mapping features into categories.
e. Implementation
The design result was then implemented by
creating a computer program using python
language and utilizing association meaning
database. Implementation could also be
Association Meaning in Identifying Pleasant Tweets
447

developed into web-based applications in order
to perform data retrieval in real time as well as
accessible to the public.
f. Testing
At this stage, testing was committed to determine
the effectiveness of systems that had been built in
classifying the tweets on Twitter. The evaluation
involved the respondents to compare the
compatibility between system output and human
judgment or expert judgment.
g. Improvements
h. Arrangement of Conclusions and Reports
4 FINDINGS AND DISSCUSION
Based on data taken from the tweets on Twitter from
users in Bandung city, below are presented some
findings of words that have experienced change into
pleasant association meaning.
Table 1: List of places with pleasant association meaning.
No
Places with
Pleasant
Association
Explanation
1
Jabar
Jabar associates as a place that
has beauty, progress, and
friendliness. Jabar is a province
that is considered advanced in
Indonesia in many ways. Jabar
associates pleasant.
2
Buah Batu
Buah Batu associates as a
pleasant place with cool air and
car free day.
3
Bandung
Bandung is the capital of Jabar
Province which is identical with
the coolness of the air, nature
tourism, and culinary. Bandung
is also known as a creative city,
a culinary paradise, and fashion
lovers paradise. Bandung
associates pleasant.
4
Braga
Braga associates as a place to
walk or hang out in Bandung
with a unique atmosphere and
classic. Braga associates
pleasant.
5
Tamansari
Tamansari associates as a
pleasant place of recreation,
because the air there is cool,
there is a zoo, also adjacent to
the famous university, ITB.
Tamansari associates pleasant.
6
Garut
Garut, a city in West Java,
identic with its natural beauty,
tourist attractions, famous for
leather products, as well as
No
Places with
Pleasant
Association
Explanation
distinctive culinary. Garut
associates as a pleasant place.
7
The Lodge
Maribaya
The Lodge Maribaya is a tourist
attraction in Lembang associated
with the natural beauty and
uniqueness of the resort. This
place associates as a pleasant
place.
Table 1 contains only an example of the analysis
that has been done. Based on the frequency of
occurrences of the word place, the words above
often appear used at pleasant tweets. From the
frequency of occurrences, the above words occupy
the top ranking compared to expressing unpleasant
meanings. Based on the rankings, the post data was
analyzed semantically based on information sources,
such as Wikipedia and based on search through
Google.
1,093 tweets from users in Bandung that have
been studied, found 37 words (adverb of place) with
pleasant association meaning, 2 words (adverb of
place) with unpleasant association meaning, and 4
words (adverb of place) which have no association.
The accuracy of the classification system that has
been built is 58.71% (using only the analysis of the
association meaning). This means the accuracy of
the system in interpreting the tweets is 58.71% in
accordance with the results of meaning manually
(human opinion).
5 CONCLUSIONS
This research is still in development process, data of
the tweets will still be added. Analysis through the
association meaning will also be developed, not
limited to the description of the place but also
equipped with semantic field analysis.
Hopefully, this research is developed completely,
either from semantic or other sides, so that it can be
used to interpret any kind of tweets. For example, it
is used to measure the happiness of the people of
Bandung and to see the response of citizens to
government policies, etc.
REFERENCES
Badan Pengembangan dan Pembinaan Bahasa,
Kementerian Pendidikan dan Kebudayaan Republik
Indonesia. 2016. Kamus Besar Bahasa Indonesia
CONAPLIN and ICOLLITE 2017 - Tenth Conference on Applied Linguistics and the Second English Language Teaching and Technology
Conference in collaboration with the First International Conference on Language, Literature, Culture, and Education
448

(online). Retrieved from: https://kbbi.kemdikbud.
go.id/
Chaer, A., 2012. Linguistik Umum (Edisi Revisi). Jakarta:
Rineka Cipta.
Chaer, A., 2002. Abdul. Pengantar Semantik Bahasa
Indonesia (Edisis Revisi). Jakarta: Rineka Cipta.
Curini, L., Iacus, S. and Canova, L., 2015. Measuring
idiosyncratic happiness through the analysis of twitter:
An application to the italian case. Social Indicators
Research, 121(2), pp.525-542.
Mitchell, L., Frank, M.R., Harris, K.D., Dodds, P.S. and
Danforth, C.M., 2013. The geography of happiness:
Connecting twitter sentiment and expression,
demographics, and objective characteristics of
place. PloS one, 8(5), p.e64417.
Parera, J. D., 2004. Teori Semantik (2
nd
ed.). Jakarta:
Penerbit Erlangga.
Pateda, M., 2010. Semantik Leksikal (2
nd
ed.). Jakarta:
Rineka Cipta.
Ponilan, I.R., Herdiani, A. and Selviandro, N., 2016.
Pengukuran Happiness Index Masyarakat Kota
Bandung Pada Media Sosial Twitter Menggunakan
Pendekatan Ontologi Top-down.
Wikipedia. 2016. Indeks Kebahagiaan. Retrieved 1
Desember 2016 from: https: // id. wikipedia. org/ wiki/
Indeks_Kebahagiaan.
APPENDIX
Below is presented attachment of the system view
that has been built.
Appendix 1. System View
Appendix 2. The Tweets
Appendix 3. Results
Association Meaning in Identifying Pleasant Tweets
449