ISSN 2087-3336 (Print) | 2721-4729 (Online)
TEKNOSAINS: Jurnal Sains, Teknologi dan Informatika
Vol. 13, No. 1, 2026, page. 1-8
http://jurnal.sttmcileungsi.ac.id/index.php/tekno
https://doi.org/10.37373/tekno.v13i1

The implementation of text mining to improve google classroom
performance based on user review
Anggita Yekti Pawestri, Annisa Uswatun Khasanah*
*

Universitas Islam Indonesia, Gedung K.H. Mas Mansyur, Kampus Terpadu UII, Jl. Kaliurang KM 14,5,
Umbulmartani, Ngemplak, Krawitan, Sleman, Daerah Istimewa Yogyakarta, 55584
* Corresponding Author: annisa.uswatun@uii.ac.id
Submitted: 23/4/2025

Revised: 15/6/2025

Accepted: 14/7/2025

ABSTRACT
The internet is developing rapidly, especially in education (e-learning). Google Classroom is an e-learning platform
that has been widely used for online learning. Google Classroom gets an average rating of 3.5 out of 5 on Google
Play, so it is necessary to research to increase the application's rating based on user reviews with text mining. The
data used in this study was 77,454 user reviews from Google Play. Data processing uses the Textblob and Naïve
Bayes Classifier methods to determine user sentiment. The sentiment analysis results with a 70/30 split data yield
an accuracy of 92.39%. Data with negative sentiments is then processed with the Association Rules method to find
words used as problem keywords, which include keywords' assignment', 'upload,' 'submit,' 'file,' 'class,'
'notification,' and 'dark.' The word is then analyzed using a fishbone diagram to find the root cause. The root of the
problems includes problems uploading files that are integrated with Google Drive, there is no setting to change
the display to 'dark mode,' The user has a good internet connection, but the file upload process is slow, and so on.
The recommendations for improvements are synchronizing file integration with Google Drive, changing the
display to dark mode according to user preferences, and updating the application server so that the file upload
process can be done more quickly.
Keywords: Sentiment analysis; text mining; Naïve Bayes classifier; Google Classroom

1.

Introduction

The rapid development of the internet has significantly influenced various aspects of society,
including education. Globally, 59.5% of the population uses the internet [1], while in Indonesia, this
figure is even higher at 79.5% [2]. This widespread internet penetration has paved the way for the
integration of digital tools in education, leading to the emergence of e-learning as a critical component
of modern teaching and learning practices. E-learning leverages information technology to facilitate
training, deliver learning materials, and enable communication between students and teachers, thereby
streamlining the learning process. One of the primary tools in e-learning is the Learning Management
System (LMS). As described by Cavus and Alhih [3], an LMS is a technological platform designed to
manage educational or training resources, distribute materials online, and monitor learning progress
effectively.
With the rapid advancement of digital technology, the use of Learning Management Systems
(LMS) in education has become increasingly widespread. Initially accelerated during the Covid-19
pandemic, online learning has now become a standard complement to traditional education. Traditional
education systems heavily relied on teachers, textbooks, and rote memorization, often leading to
monotonous and disengaging learning experiences for students. Modern technology-enhanced learning,
however, complements the educational process by accommodating diverse learning styles and abilities,
ultimately enhancing student engagement and improving learning outcomes [4]. The integration of
TEKNOSAINS: Jurnal Sains, Teknologi & Informatika is licensed under a Creative
Commons Attribution-NonCommercial 4.0 International License. ISSN 2087-3336
(Print) | 2721-4729 (Online)

2 Anggita Yekti Pawestri, Annisa Uswatun Khasanah
The implementation of text mining to improve google classroom performance based on user
review

technology, communication, and information in education has had a profound impact [5]. Key trends in
learning include (1) a shift from teacher-centered to student-centered approaches, (2) the increasing
popularity and expansion of open and distance learning, and (3) the enhanced accessibility to a wider
variety of learning resources.
Even in the post-pandemic era, online and hybrid learning models continue to play a vital role in
education. Schools and universities leverage LMS platforms not only to provide remote learning
opportunities but also to enhance in-class experiences. Many distance learning activities have
collaborated with online learning platforms. Google Classroom is one of the most widely used learning
platforms. Based on a survey conducted by Arus Survei Indonesia (ASI), 26.1% of people have used
the Google Classroom platform during distance learning [6]. Google Classroom occupies the highest
percentage as an application for distance learning activities, followed by other platforms such as
Ruangguru (17.1%) and Rumah Belajar (15.2%). Platforms like Google Classroom remain popular due
to their user-friendly interface and integration with other digital tools, enabling educators to efficiently
manage assignments, distribute materials, and foster communication with students. This trend signifies
a shift toward a more technology-driven approach in education, ensuring accessibility and innovation in
learning.
The increasing number of users of Google Classroom has raised criticism and suggestions from
users. Based on the rating level expressed by users on Google Play in 2024, the Google Classroom
application received an average rating of 3.5 out of 5 in the Google Play store. The rating is relatively
low compared to other Google applications, so improvements need to be made to increase the rating of
the Google Classroom application. Before enhancements are made, it is necessary to find the root of the
problem that causes the Google Classroom application to have a low rating. Reviews from Google
Classroom app users are used as a source of data to explore issues from the app. Text mining is one
technique that can be implemented to classify data by finding interesting patterns from large text data
sets [7]. According to Antons et al. [8], text mining requires a quantitative approach to analyze text data.
In this research, text mining is used to explore the review data of Google Classroom application
users, especially on the Android platform. Reviews that become data used in this research are in English.
The language was chosen because it is the majority language used in giving reviews, so that it can
represent criticism from users. Data collection will be focused on the Google Play Store platform by
scraping data through the AppFollow website. In this research, sentiment classification analysis is
carried out to classify reviews with positive and negative sentiments. The Naïve Bayes Classifier method
is used in this research because it is considered one of the sentiment analysis methods that are simple,
easy to understand, and have a reasonably high accuracy rate [9]. This method can project possibilities
that can happen in the future based on experiences that have occurred, which is now known as Bayes'
Theorem. In addition, a lexicon-based labeling library in Python, Textblob, will also be used. Textblob
is one of the libraries in Python for text data processing and uses NLTK for Natural Language Processing
(NLP), which is unlabeled [10]. The sentiment analysis classification is a hybrid method between
machine learning and lexicon-based sentiment analysis. Association rule mining will also be used in this
research. According to Santoso [11], association rules are used to find rules or combinations of items
that occur most frequently in data. This study uses association rules to find combinations of these words.
After being classified, the root of the problem can be found using root cause analysis in the form of a
fishbone diagram so that it can be used as a reference for developing recommendations for
improvements to the Google Classroom application.
2.

Method

Text mining has been widely used for sentiment analysis in user reviews. Conducted sentiment
analysis on 66,000 MOOC reviews using text mining techniques [12]. The study compared machine
learning, ensemble learning, and deep learning methods for sentiment classification in education.
Results showed that deep learning architectures achieved the highest predictive accuracy among the
methods tested. Conducted a sentiment analysis about online education during the Covid-19 era from
Twitter [13]. This research aims to evaluate the effectiveness of e-learning by examining people's
sentiments toward it. Integrated text mining and machine learning to analyze student evaluations and
uncover how sentiments, emotions, and gender influence teacher recommendation using the proposed
model, the EPDM + ML [14]. The model achieved perfect prediction accuracy, highlighting its
effectiveness in improving teaching quality through data-driven insights. Conducted study in sentiment
analysis for TikTok using Naive Bayes and Support Vector Machine [15].

ISSN 2087-3336 (Print) | 2721-4729 (Online)
DOI 10.37373/tekno.v13i1.1479

3

This current study applied text mining to determine user sentiment in Google Classroom. The data
used in this study consists of 77454 user reviews. This data was obtained using web scraping through
the AppFollow site for six months. The data preprocessing stage is carried out to clean the data of
unnecessary information and anomalies. This process was carried out using Jupyter Notebook and
Python 3.9. The data preprocessing includes several steps:
a.

Case folding: Homogenizes all letters so that letters that were initially capitalized become
lowercase in Table 1.
Table 1. Case folding step

Before Case Folding
This app is mostly like trash 🤦â€â™€ï¸ I've
been using this app for a while and i changed my
acc password, and just like that, i can't log out
and whenever I use it, it will tell me something
went wrong and wants me to try again…
b.

After Case Folding
this app is mostly like trash 🤦â€â™€ï¸ i've
been using this app for a while and i changed my
acc password, and just like that, i can't log out
and whenever i use it, it will tell me something
went wrong and wants me to try again…

Cleaning: remove punctuation marks to delete unwanted symbols or punctuation marks in text data,
remove non-ASCII characters, which are Character Encoding Standard for Electronic
Communication. And the last was removing emojis in the text data to Table 2, Table 3, Table 4,
Table 5

Table 2. Punctuation removal step
Before Remove Punctuation
After Remove Punctuation
this app is mostly like trash 🤦â€â™€ï¸ i've this app is mostly like trash 🤦â€â™€ï¸ ive
been using this app for a while and i changed my been using this app for a while and i changed my
acc password, and just like that, i can't log out and acc password and just like that i cant log out and
whenever i use it, it will tell me something went whenever i use it it will tell me something went
wrong and wants me to try again...
wrong and wants me to try again
Table 3. Non-ASCII character removal step
Before Non-ASCII Character Removal
After Non-ASCII Character Removal
this app is mostly like trash ðÿ¤¦â€â™€ï¸ ive this app is mostly like trash ive been using this
been using this app for a while and i changed my app for a while and i changed my acc password
acc password and just like that i cant log out and and just like that i cant log out and whenever i
whenever i use it it will tell me something went use it it will tell me something went wrong and
wrong and wants me to try again
wants me to try again
Table 4. Emoji removal step
Before Emoji Removal
After Emoji Removal
this app is mostly like trash ive been using this app this app is mostly like trash ive been using this
for a while and i changed my acc password and app for a while and i changed my acc password
just like that i cant log out and whenever i use it it and just like that i cant log out and whenever i
will tell me something went wrong and wants me use it it will tell me something went wrong and
wants me to try again
to try again
c.

Tokenizing: separates the words in a sentence, leaving only the needed words.

Table 5. Tokenizing step
Before Tokenizing
After Tokenizing
this app is mostly like trash ive been using this [‘this’, ‘app’, ‘is’, ‘mostly’, ‘like’, ‘trash’, ‘ive’,
app for a while and i changed my acc password ‘been’, ‘using’, ‘this’, ‘app’, ‘for’, ‘a’, ‘while’,
and just like that i cant log out and whenever i ‘and’, ‘i’, ‘changed’, ‘my’, ‘acc’, ‘password’,
use it it will tell me something went wrong and ‘and’, ‘just’, ‘like’, ‘that’, ‘i’, ‘cant’, ‘log’, ‘out’,
wants me to try again
‘and’, ‘whenever’, ‘use’ ‘it’, ‘it’, ‘will’, ‘tell’,
‘me’, ‘something’, ‘went’, ‘wrong’, ‘and’,
‘wants’, ‘me’, ‘to’, ‘try’, ‘again’]

4 Anggita Yekti Pawestri, Annisa Uswatun Khasanah
The implementation of text mining to improve google classroom performance based on user
review

d.

Stopwords removal: remove meaningless words, including the, and, but, and or, and others to Table
6.

Table 6. Stopwords removal step
Before Tokenizing
After Tokenizing
[‘this’, ‘app’, ‘is’, ‘mostly’, ‘like’, ‘trash’, ‘ive’, ['app', 'mostly', 'like', 'trash', 'using', 'app',
‘been’, ‘using’, ‘this’, ‘app’, ‘for’, ‘a’, ‘while’, 'changed', 'acc', 'password', 'like', 'log',
‘and’, ‘i’, ‘changed’, ‘my’, ‘acc’, ‘password’, 'whenever', 'use', 'tell', 'something', 'went',
‘and’, ‘just’, ‘like’, ‘that’, ‘i’, ‘cant’, ‘log’, ‘out’, 'wrong', 'wants', 'try', 'again']
‘and’, ‘whenever’, ‘use’ ‘it’, ‘it’, ‘will’, ‘tell’,
‘me’, ‘something’, ‘went’, ‘wrong’, ‘and’,
‘wants’, ‘me’, ‘to’, ‘try’, ‘again’]
After preprocessing, the data label must be given. Textblob, one of the Python libraries that
performs sentiment analysis with a lexicon-based approach, provides labels on the data. The sentiment
analysis results with Textblob will give the output of subjectivity and polarity values. The polarity value
is used as a reference to determine the sentiment of text data into the negative, neutral, or positive
category. Labeling criteria on Textblob refer to the polarity score with the following conditions [16]:
a. Polarity score < 0, then the sentiment is negative
b. Polarity score = 0, then neutral sentiment
c. Polarity score > 0, then positive sentiment
Based on the results of sentiment labeling using the Textblob library, the data was divided into
negative sentiment (27,233 data), positive sentiment (27,931 data), and neutral sentiment (22,290 data).
The visualization of the percentage of sentiment labeling results is shown in Figure 1. It can be known
that positive and negative sentiments have similar percentages, which means from the Google
Classroom data review the number of negative comments and positive comments were almost the same.

Figure 1. Visualization of sentiment labeling results
3. Result and Discussion
a.

Classification Using Naïve Bayes Algorithm
Classification with the Naïve Bayes algorithm is done to determine the accuracy of the
classification model in predicting sentiment. The Naïve Bayes Classifier algorithm is a machine learning
model that is supervised or needs to be trained. The illustration of the classification model is shown in
Figure 2.

Figure 2. Classification model
Therefore, training and test data are required to apply the classification model. Data splitting uses
the 'train_test_split' library from sklearn in Python. Several combinations of training and testing data
were generated to get the best accuracy; the combinations are: 50/50, 60/40, 70/30, and 80/20 [17].

ISSN 2087-3336 (Print) | 2721-4729 (Online)
DOI 10.37373/tekno.v13i1.1479

5

Training data was used to build the model, and test data was used to evaluate the model. The
performance measurement was performed by calculating the model accuracy and F1 score. The
comparative results of classification using the Naïve Bayes classifier with four combinations of split
data are shown in Table 7.
Table 7. Comparison of split data size, accuracy, and f1 score
Accuracy
F1 Score
Split Data 50/50
92.19%
92.06%
60/40
92.27%
92.3%
70/30
92.39%
92,.5%
80/20
92.33%
92.17%
Based on the accuracy rate and F1 score values obtained, it is known that classification with a data
splitting size of 70% training data and 30% test data produces the best output with an accuracy value of
92.39%.
b.

Association Rules
The data extracted is 27,233 customer reviews with negative sentiments containing criticisms and
complaints about the Google Classroom application. In data processing, the stemming process is
implemented using the Snowball Stemmer algorithm, which aims to minimize word diversity. The
Snowball Stemmer algorithm is a development of the Porter Stemmer algorithm, which is considered to
have the best output compared to other stemmer algorithms [18]. Association rules processing uses the
FP-Growth algorithm. The FP-Growth algorithm is an efficient algorithm for finding frequent item sets,
which are considered to have better performance when compared to the Apriori algorithm [19]. Data
mining is done to find the most frequent itemset in the data with FP-Growth by creating an FP-Tree
structure) [20].
Association rules data with the FP-Growth algorithm is applied in RapidMiner with min support of
0.001 and min confidence of 0.1. The support value represents the percentage of words in the data, while
the confidence value represents the strength of the association between the words formed in Table 8. If
the resulting lift ratio value is more than 1, it can be concluded that the relationship or association
between the conclusion and premises is significant. The words which are frequently found in data are
'assignment', 'upload', 'submit', 'file', 'class', 'notification', and 'dark'.
Table 8. Frequent words
Words
Support
class
0.042
assignment
0.036
upload
0.024
submit
0.019
file
0.018
notification
0.017
dark
0.009
Based on the AR results, it can be concluded that the frequent words are correlated with several
other words to determine the specific issues. The assignment was correlated with miss, late, post,
notification, etc. Upload was correlated with slow, speed, assignment, etc. Submit was correlated with
time, assignment, worst, late, take, miss, etc. File was correlated with attachment, time, take, upload,
slow, etc. Class was correlated with students, teacher, online, attend, worst, etc. Notifications ware
correlated with received, late, missed, assignment, etc. Dark was correlated with the mode, please,
theme.
c.

Recommendations for Improvement
Based on the data processing results with association rules, the problems that can be analyzed
related to assignments; upload and submit files; class entry system; notifications issues; and displays
problems. The root cause is analyzed based on four aspects: product, people, place, and process. The
product aspect is related to the product feature, which is the Google Classroom application. The people
aspect is associated with the human error problems related to obstacles in using the application. The

6 Anggita Yekti Pawestri, Annisa Uswatun Khasanah
The implementation of text mining to improve google classroom performance based on user
review

place aspect refers to difficulties in accessing the Google Classroom application. The last aspect is the
process related to the problem users encounter when running the application. After the data has been
analyzed, the improvement recommendation is formulated. The recommendations based on four aspects
are shown in Table 9, Table 10, Table 11, Table 12.
Table 9. Recommendations for improvement according to product aspects
Product Aspects
Problems
Recommendations for Improvement
1. Obstacles in uploading files integrated
1. Improve file integration with Google Drive
with Google Drive
2. Add a feature to change the interface theme of
2. No theme variation for the application
the Google Classroom application to make it
interface
more attractive to users
3. No setting to change the display to 'dark 3. Settings to change the display to dark mode
mode'
according to the user's preference
Table 10. Recommendations for improvement according to the people aspect
People Aspects
Problems
Recommendations for Improvement
Communication between teachers and students is Improve and enhance the communication forum
not optimal, leading to miscommunication,
between students and teachers.
Table 11. Recommendations for improvement according to the place aspect
Place Aspects
Problems
Recommendations for Improvement
1. There is no notification even though the user 1. Optimize notifications so users can still
has an internet connection
receive notifications and do not experience
2. Users have a good internet connection, but
delays
the file uploading process is slow
2. Update the Google Classroom server so the
file upload process can be done faster

3.
4.
5.
6.
7.
8.

Table 12. Recommendations for improvement according to the process aspect
Process Aspects
Problems
Recommendations for Improvement
The assignment is submitted on time, but the
1. Improving the assignment assessment feature
status is late
2. Improvements need to be made to the
Assignments that have been collected have a
submission process for various types of files
missed status on the Dashboard
3. Improvements need to be made to the server
There is an error in the grading feature
4. Updates are needed to resolve bugs.
Problems collecting photo and video file types
The result of the uploaded file is not visible
Failed to enter the classroom
Late notification
No notification when there is a new comment

4.

Conclusion

1.
2.

Based on the study that has been conducted, it can be concluded that the best combination of
splitting data to build the classification model is 70/30. The accuracy rate using Naïve Bayes Classifier
is 92.39%. Based on the processing of association rules with the FP-Growth method, it can be found
that the word appears most often when the support value of the word is examined. The words that appear
most often in Google Classroom application user reviews are 'class', 'assignment', 'upload', 'submit', 'file',
'notification', and 'dark'. The root of the problem can be formulated using a fishbone diagram based on
four aspects: product, people, place, and process. The improvement recommendations are developed to
increase the performance of Google Classroom.
Reference

ISSN 2087-3336 (Print) | 2721-4729 (Online)
DOI 10.37373/tekno.v13i1.1479

[1]
[2]

[3]

[4]

[5]

[6]

[7]

[8]

[9]

[10]
[11]

[12]

[13]

[14]

[15]

[16]

7

A. Petrosyan, “Number of internet and social media users worldwide as of October 2024,”
https://www.statista.com/statistics/617136/digital-population-worldwide/.
Asosiasi Penyelenggara Jasa Internet Indonesia, “APJII Jumlah Pengguna Internet Indonesia
Tembus 221 Juta Orang,” https://apjii.or.id/berita/d/apjii-jumlah-pengguna-internet-indonesiatembus-221-jutaorang#:~:text=Asosiasi%20Penyelenggara%20Jasa%20Internet%20Indonesia%20(APJII)%20
mengumumkan%20jumlah%20pengguna%20internet,jiwa%20penduduk%20Indonesia%20tah
un%202023.
N. Cavus and M. S. Alhih, “Learning Management Systems Use in Science Education,” Procedia
Soc Behav Sci, vol. 143, pp. 517–520, Aug. 2014, doi: 10.1016/j.sbspro.2014.07.429.
https://doi.org/10.1016/j.sbspro.2014.07.429
D. Ambarwati, U. B. Wibowo, H. Arsyiadanti, and S. Susanti, “Studi literatur: Peran inovasi
pendidikan pada pembelajaran berbasis teknologi digital,” Jurnal Inovasi Teknologi Pendidikan,
vol. 8, no. 2, pp. 173–184, 2022.
R. Dwi Gunawan, A. Sutisna, and E. Fitri Ana, “Literature review: The role of learning
management system (LMS) in improving the digital literacy of educators,” Jurnal Inovasi
Teknologi Pendidikan, vol. 11, no. 2, pp. 116–123, 2024, doi: 10.21831/jitp.v11i2.56326.
I. Kamil and B. Galih, “Survei: Google Classroom Jadi Platform Belajar Paling Sering
Digunakan Saat PJJ,” https://nasional.kompas.com/read/2020/10/16/18264341/survei-googleclassroom-jadi-platform-belajar-paling-sering-digunakan-saat?page=all#google_vignette.
E. E. Pratama and R. L. Atmi, “A Text Mining Implementation Based on Twitter Data to Analyse
Information Regarding Corona Virus in Indonesia,” Journal of Computers for Society, vol. 1, no.
1, pp. 91–100, 2020, [Online]. Available: https://ejournal.upi.edu/index.php/JCS
D. Antons, E. Grünwald, P. Cichy, and T. O. Salge, “The application of text mining methods in
innovation research: current state, evolution patterns, and development priorities,” in R and D
Management, Blackwell Publishing Ltd, Jun. 2020, pp. 329–351. doi: 10.1111/radm.12408.
https://doi.org/10.1111/radm.12408
P. S. M. Suryani, L. Linawati, and K. O. Saputra, “Penggunaan Metode Naïve Bayes Classifier
pada Analisis Sentimen Facebook Berbahasa Indonesia,” Majalah Ilmiah Teknologi Elektro, vol.
18,
no.
1,
p.
145,
May
2019,
doi:
10.24843/mite.2019.v18i01.p22.
https://doi.org/10.24843/MITE.2019.v18i01.P22
C. Kaur and A. Sharma, “Twitter Sentiment Analysis on Coronavirus using Textblob,”
EasyChair Preprint, no. 2974, 2020.
M. H. Santoso, “Application of Association Rule Method Using Apriori Algorithm to Find Sales
Patterns Case Study of Indomaret Tanjung Anom,” Brilliance: Research of Artificial
Intelligence, vol. 1, no. 2, pp. 54–66, Dec. 2021, doi: 10.47709/brilliance.v1i2.1228.
https://doi.org/10.47709/brilliance.v1i2.1228
A. Onan, “Sentiment analysis on massive open online course evaluations: A text mining and
deep learning approach,” Computer Applications in Engineering Education, vol. 29, no. 3, pp.
572–589, May 2021, doi: 10.1002/cae.22253. https://doi.org/10.1002/cae.22253
M. Mujahid et al., “Sentiment analysis and topic modeling on tweets about online education
during covid-19,” Applied Sciences (Switzerland), vol. 11, no. 18, Sep. 2021, doi:
10.3390/app11188438. https://doi.org/10.3390/app11188438
K. Okoye, A. Arrona-Palacios, C. Camacho-Zuñiga, J. A. G. Achem, J. Escamilla, and S.
Hosseini, “Towards teaching analytics: a contextual model for analysis of students’ evaluation
of teaching through text mining and machine learning classification,” Educ Inf Technol (Dordr),
vol. 27, pp. 3891–3933, 2022. https://doi.org/10.1007/s10639-021-10751-5
Friska Aditia Indriyani, Ahmad Fauzi, and Sutan Faisal, “Analisis sentimen aplikasi tiktok
menggunakan algoritma naïve bayes dan support vector machine,” TEKNOSAINS : Jurnal Sains,
Teknologi dan Informatika, vol. 10, no. 2, pp. 176–184, Jul. 2023, doi:
10.37373/tekno.v10i2.419. https://doi.org/10.37373/tekno.v10i2.419
P. Shah, “Sentiment Analysis using TextBlob,” https://towardsdatascience.com/my-absolute-goto-for-sentiment-analysis-textblob-3ac3a11d524.

8 Anggita Yekti Pawestri, Annisa Uswatun Khasanah
The implementation of text mining to improve google classroom performance based on user
review

[17]

[18]
[19]

[20]

A. Rácz, D. Bajusz, and K. Héberger, “Effect of dataset size and train/test split ratios in qsar/qspr
multiclass
classification,”
Molecules, vol.
26,
no. 4,
Feb. 2021,
doi:
10.3390/molecules26041111. https://doi.org/10.3390/molecules26041111
M. Anjali and G. Jivani, “A Comparative Study of Stemming Algorithms,” Int.J. Comp. Tech.
Appl, vol. 2, no. 6, pp. 1930–1938, 2011, [Online]. Available: www.ijcta.com
M. S. Mythili and A. R. M. Shanavas, “Performance Evaluation of Apriori and FP-Growth
Algorithms,” Int J Comput Appl, vol. 79, no. 10, pp. 975–8887, 2013.
https://doi.org/10.5120/13779-1650
D. Widiastuti and N. Sofi, “Analisis Perbandingan Algoritma Apriori Dan Fp-Growth Pada
Transaksi Koperasi,” UG Journal, vol. 8, no. 1, 2014.