Computer Science and Information Technologies

Vol. 5, No. 3, November 2024, pp. 306~314

ISSN: 2722-3221, DOI: 10.11591/csit.v5i3.pp306-314  306

Vector space model, term frequency-inverse document

frequency with linear search, and object-relational mapping
Django on hadith data search

Ichsan Taufik, Agra, Yana Aditia Gerhana

Department of Informatics, Faculty Science and Technologi, UIN Sunan Gunung Djati, Bandung, Indonesia

Article Info ABSTRACT

Article history: For Muslims, the Hadith ranks as the secondary legal authority following the
Quran. This research leverages hadith data to streamline the search process
Received Jun 2, 2024 within the nine imams’ compendium using the vector space model (VSM)
Revised Oct 8, 2024 approach. The primary objective of this research is to enhance the efficiency
Accepted Oct 19, 2024 and effectiveness of the search process within Hadith collections by
implementing pre-filtering techniques. This study aims to demonstrate the
potential of linear search and Django object-relational mapping (ORM)
Keywords: filters in reducing search times and improving retrieval performance, thereby
facilitating quicker and more accurate access to relevant Hadiths. Prior
Hadith 9 imams studies have indicated that VSM is efficient for large data sets because it
Linear search assigns weights to every term across all documents, regardless of whether
Orm django they include the search keywords. Consequently, the more documents there
Search are, the more protracted the weighting phase becomes. To address this, the
Vector space model current research pre-filters documents prior to weighting, utilizing linear
search and Django ORM as filters. Testing on 62,169 hadiths with 20
keywords revealed that the average VSM search duration was 51 seconds.
However, with the implementation of linear and Django ORM filters, the
times were reduced to 7.93 and 8.41 seconds, respectively. The recall@10
rates were 79% and 78.5%, with MAP scores of 0.819 and 0.814,
accordingly.
This is an open access article under the CC BY-SA license.

Corresponding Author:
Yana Aditia Gerhana
Department of Informatics, Faculty Science and Technology, UIN Sunan Gunung Djati Bandung
Jl. A. H. Nasution 105, Bandung 40614 Indonesia
Email: yanagerhana@uinsgd.ac.id

1. INTRODUCTION
Information retrieval (IR), according to Salton in [1], is a system that can automatically retrieve
information according to user needs from a set of data. One of the classic methods that is popular and has
been tested in IR is the vector space model (VSM) [2]–[4]. VSM represents keywords in each document as a
vector. The similarity values of the keywords and documents are then obtained from the angles formed by the
two vectors.
In online news classification research [5], testing of the VSM method managed to achieve a high
level of accuracy, namely 91.25% in four tests. The test was conducted by dividing the research object into
training data and testing data. Other research showing the success of VSM in retrieving information can be
found in journals [6]–[8].
To get the similarity value of two vectors in VSM [4], we first need to calculate the weight of each
vector using the term weighting scheme. Methods used in word weighting include simple term frequency-

Journal homepage: http://iaesprime.com/index.php/csit

Comput Sci Inf Technol ISSN: 2722-3221  307

inverse document frequency (TF-IDF), incremental TF-IDF, phrase-augmented TF-IDF, latent semantic
index (LSI), and document-to-vector (D2V). A comparison of these methods is presented in a comparative
study [9] for the case of text similarity to patent data managed by the United States Patent and Trademark
Office (USPTO). For this research context, it is known that simple TF-IDF is a more suitable method because
the objects used do not require complicated natural language processing (NLP) processes.
Furthermore, in the TF-IDF [5] algorithm, the weight value for each vector is obtained by
combining the frequency of occurrence of a word in a document with the frequency of the inverse document
(IDF) that contains the word. After that, we can rank the documents that have the greatest degree of similarity
to keywords using text similarity calculations using the Cosine Similarity, Jaccard Similarity, or Euclidean
Distance approaches. A Comparison of these three approaches [10] identifying top news items on news sites
and measuring the similarity between the same two news items in two different languages based on the same
event found that Cosine Similarity has the highest accuracy value compared to the other approaches.
Django is a web framework written in the Python programming language, rich in libraries [11]–[13].
These libraries facilitate the implementation of the VSM with tools for stopword removal, literary, numpy,
and punkt, which are used for preprocessing -- removing meaningless words and punctuation marks before
the TF-IDF calculation. Additionally, Django also supports ORM techniques, simplifying database queries.
Previous studies have highlighted that the VSM method tends to require less time to process
searches with large amounts of data. In one study [14], the average search took 2.24 seconds, with a success
rate of 81% on 75 hadith data entries. Another study [15] took 6.042 seconds, even after incorporating the
Hamming distance algorithm to expedite the search.
This is caused by the VSM process of assigning weights to each word in all documents, even if the
document does not contain the keyword being searched for. Thus, the more documents used for searching,
the longer the weighting process will be. However, from the studies conducted, the data generated from
searches using the VSM method is certain to contain at least one word of the keyword. Therefore, in this
study, document filtering was carried out prior to the weighting process.
Linear search, also commonly known as sequential search, is a fairly simple and easy search
algorithm. It works by comparing each element in succession, starting with the first element, until the sought
element is found or all elements have been examined [16]. In contrast to Binary Search, which can only work
if the data has been sorted (requiring a sorting process before the search can be carried out).
By using Linear Search for the filtering process, it can be ensured that the document to be weighted
contains at least one word of the searched keyword. Additionally, Linear Search does not require prior
sorting, saving time in the search process. Likewise, ORM simplifies making search queries in the filtering
process, using linear search as a filter comparison. A similar study [17] explores the use of TF-IDF and VSM
for efficient retrieval of medical records. However, unlike our research which focuses on performance
improvements using linear and Django ORM filters, this study emphasizes the application of TF-IDF and
VSM in medical document retrieval. Based on 1,000 medical record documents and tested with 20 search
queries, the results showed an average precision value of 0.548 and an average recall value of 0.796.
The research object used is hadith data from the Hadith Encyclopedia of the Book of 9 Imams in the
LIDWA Pusaka application. This application contains approximately 62,000 hadiths, which can be used as
learning media or research objects [18]. Hadith is the second source of law for Muslims after the Quran. The
use of hadith data [2] as an object of research is intended to make it easier for users who want to search
hadith in the Book of Nine Imams by applying the VSM concept. As already explained, the advantage of
VSM is its ability to return the most relevant information for the keywords you are searching for, which can
be in the form of words, phrases or sentences. Currently, most hadith data searches can only be done using
words or the concept of string matching, making it difficult for users to find the hadith in question.
The goal of this research is to improve the search function for hadiths in the LIDWA Pusaka app
database using the VSM, TF-IDF with linear search, and ORM Django. This method helps users find relevant
hadiths more easily by allowing searches with words, phrases, or sentences, unlike traditional methods that
only use exact word matches. This makes it easier for users to access the hadiths they need for learning or
research.

2. METHOD
2.1. VSM
In the journal [19], Anna and Hendini explained that the VSM is a method or algorithm that is often
used for an information retrieval system. This algorithm is a model that is used to measure the similarity
(Similarity terms) between a document and a query by means of term weighting. In the information retrieval
system, similarities between documents are defined based on a bag-of-words representation and converted to
a VSM. The relevance of a document to a query is based on the similarity between the document vector and
the query vector, as seen in Figure 1.
Vector space model, term frequency-inverse document frequency with linear search…(Ichsan Taufik)
308  ISSN: 2722-3221

Figure 1. VSM illustration [20]

𝑞.𝑑𝑓𝑗 ∑𝑡𝑖=1 𝑤𝑖𝑞 .𝑤𝑖𝑗

𝑠𝑖𝑚(𝑞, 𝑑𝑗 ) = |𝑞| = (1)
.|𝑑𝑗 | 2 2
√∑𝑡𝑗=1(𝑤𝑖𝑞 )2 . √∑𝑡𝑖=1(𝑤𝑖𝑗 )2

Information:
Sim (di, dj) = Similarity between query and document
||D1|| = Document vector length 1
||D2|| = Document vector length 2
Wij = Term weight in the document
Wiq = Query weight in the document
The basic concept of the VSM is to calculate the distance between documents and then sort them
based on their closeness. The workings of the VSM begin with case folding, stopword removal, stemming,
and tokenizing, which are stages of cutting the input string based on each word and breaking the document
into word frequency tables. All the words in the document are combined into one, which is called a term.
Each document is displayed as a vector, which will be compared with the terms that have been formed.
Similarity analysis to measure the similarity of documents is done by calculating the cosine [21] of the
distance between the documents.

2.2. Linear search

Linear search is one of the most popular and simple search algorithms [22], [23]. The search
technique involves scanning data sequentially from start to finish based on the desired keywords. A
significant advantage of this algorithm is that if the target data is located near the beginning, it can be found
quickly. Conversely, a major disadvantage is that if the target data is near the end, the search can take a long
time [24].

2.3. ORM
ORM is a programming method used to convert data from an object-oriented programming (OOP)
language environment to a relational database environment [25], [26]. The advantages of using ORM
include,
− It can speed up program development, such as query repetition, because tables are represented as
objects.
− Data access becomes more abstract and portable based on the database vendor.
− It supports encapsulating business rules at the data access layer.
− It is able to generate boilerplate code for basic CRUD functions.

2.4. System analysis

The system to be built is a website-based application using Django [12] as the framework. The
system architecture is shown in Figure 2. The client or user makes an HTTP request for a hadith search via a
browser, which is then forwarded to a web server such as Apache, IIS, Nginx, Tomcat, and so on. The server
processes the request and returns it as an HTTP response according to the user's request (in this research, it is
a list of hadiths). Figure 3 shows the system flow to be developed in this study.

2.5. Confussion matrix

The confusion matrix is a crucial tool in evaluating of classification models in machine learning. It
provides a detailed breakdown of the model’s performance by comparing actual and predicted classifications.
The matrix consists of four key components: True positives (TP), false positives (FP), true negatives (TN),

Comput Sci Inf Technol, Vol. 5, No. 3, November 2024: 306-314

Comput Sci Inf Technol ISSN: 2722-3221  309

and false negatives (FN). These components help calculate various performance metrics such as accuracy,
precision, recall, and F1 score, which are essential for understanding a model’s strengths and weaknesses.
For instance, in a medical diagnosis scenario, a high recall indicates the model’s effectiveness in identifying
patients with a disease, while precision highlights its ability to avoid false alarms [27].

Figure 2. System arhitecture

Figure 3. System flowchart

3. RESULT AND DISCUSSION

3.1. Implementation
In its implementation, this website has three interface pages, including the home page, search page,
and detail page as seen in Figures 4 to 6. The home page serves as the main entry point, providing an
overview and easy navigation. The search page allows users to find specific content quickly and efficiently.
The detail page offers in-depth information and additional resources related to the selected content.
Vector space model, term frequency-inverse document frequency with linear search…(Ichsan Taufik)
310  ISSN: 2722-3221

Figure 4. Home page

Figure 5. Search page

Figure 6. Detail page

Comput Sci Inf Technol, Vol. 5, No. 3, November 2024: 306-314

Comput Sci Inf Technol ISSN: 2722-3221  311

On the search page, this algorithm is applied. Figures 7 to Figure 15 are details of the programming
syntax using the Python language. These figures illustrate the step-by-step implementation of the algorithm.
Additionally, they highlight key functions and methods used to achieve the desired search functionality.

a) Function gets documents and tokenizing

Figure 7. Get documents and tokenizing VSM

Figure 8. Get documents and tokenizing VSM with filter ORM

Figure 9. Get documents and tokenizing with filter linear search

Figure 10. Linear search function

b) Term frequency

Figure 11. Term frequency

Inverse document frequency

Figure 12. IDF without filter (VSM only)

Vector space model, term frequency-inverse document frequency with linear search…(Ichsan Taufik)
312  ISSN: 2722-3221

Figure 13. IDF with filter (linear or ORM)

c) Ranking of the 100 highest data

Figure 14. Ranking of the 100 highest data

d) TF-IDF and cosine similarity

Figure 15. TF-IDF and cosine Similarity

3.2. Testing
Algorithm testing was conducted to determine the performance of the VSM algorithm after adding a
filter by using the recall@k (recall top k), precision@k (precision top k), and mean average precision (MAP)
approaches with a value of k=10 to the performance of the VSM algorithm without filters. In this research,
testing was carried out using 20 keywords in the form of words, phrases, and sentences for each filter
method. The use of TF-IDF in testing results transformed textual data into numerical values, highlighting the
importance of each term within the dataset. The TF-IDF values were crucial in creating document vectors for
the VSM, which were then used to compute similarities between documents and search queries, resulting in
performance metrics such as recall@k, precision@k, and MAP. The value of k=10 was chosen to evaluate
the top 10 results returned by the VSM algorithm, providing practical insights into the algorithm’s
effectiveness in returning the most relevant results within a manageable number of top results. The selection
of 20 keywords ensured a diverse set of search queries, including words, phrases, and sentences, to
comprehensively evaluate the algorithm’s performance across different query complexities and lengths,
ensuring robustness and generalizability of the results.
Algorithm testing with recall@10 using Linear Search and ORM filters is valued 79%. This means
that about 8 of the top 10 hadiths predicted by VSM appear in searches using the filter in the top 10 orders.
MAP evaluates the average precision across several keywords, with a value of 0.814 for the VSM method
with the ORM filter and 0.819 for the VSM method with Linear Search filter. It can be concluded that both
methods work well in terms of sorting.

4. CONCLUSION
From the results of the Black Box testing that has been carried out, the system search using the
VSM method, both with the Linear Search filter and the Django ORM, has been successfully implemented

Comput Sci Inf Technol, Vol. 5, No. 3, November 2024: 306-314

Comput Sci Inf Technol ISSN: 2722-3221  313

and works well functionally, as expected. The testing confirmed that the search functionality is robust and
efficient. Additionally, the integration with Django ORM ensures seamless database interactions. Overall, the
system meets the desired performance criteria and user expectations. VSM performance in this research was
carried out with 20 sample queries. For searches using the VSM method alone, an average time of 51 seconds
was obtained with 62,169 hadith documents. This is quite long for a search method. Therefore, VSM with
filters can be used as an option. The filter here is applied before carrying out similarity calculations because it
can reduce the accumulation of words during weighting, thus speeding up the process. Tests such as
recall@k, precision@k, and MAP tests were also carried out to assess the relevance of the relevance of the
algorithm after applying the filter. The results show that adding filters to VSM can work well and is relevant.
Data filters can be developed further by exploring other word filter algorithms. Additionally, the VSM in this
research performs exceptionally well for searches using keywords in words, phrases, or sentences. However,
it does not include searches using synonyms for these keywords. This can be addressed through a query
expansion approach to increase the relevance of search results or by other methods.

ACKNOWLEDGEMENTS
The authors thank the Center for Research and Publications, Institute for Research and Community
Service at UIN Sunan Gunung Djati Bandung. In most cases, sponsor and financial support
acknowledgments.

REFERENCES
[1] M. Mayeni, W. W. Winarno, and A. Sunyoto, “Information retrieval of thesis documents to determine their similarity to existing
research" (Indoensian) "Information retrieval dokumen tesis untuk mengetahui kemiripannya dengan penelitian yang telah ada,”
Jurnal TRANSFORMASI, vol. 12, no. 2, pp. 105–115, 2016.
[2] S. E. Pratama, W. Darmalaksana, D. Sa’adillah Maylawati, H. Sugilar, T. Mantoro, and M. A. Ramdhani, “Weighted inverse
document frequency and vector space model for hadith search engine,” Indonesian Journal of Electrical Engineering and
Computer Science, vol. 18, no. 2, pp. 1004–1014, 2020, doi: 10.11591/ijeecs.v18.i2.pp1004-1014.
[3] A. Fauzi and G. Ginabila, “Online news classification using TF-IDF weighting and cosine similarity,” “Online news classification
using TF-IDF weighting and cosine similarity,” (Indonesian) “Information retrieval system pada file pencarian dokumen tesis
berbasis text menggunakan metode vector space model,” Jurnal Pilar Nusa Mandiri, vol. 15, no. 1, pp. 41–46, 2019, doi:
10.33480/pilar.v15i1.61.
[4] J. E. Dobson, “Vector hermeneutics: On the interpretation of vector space models of text,” Digital Scholarship in the Humanities,
vol. 37, no. 1, pp. 81–93, 2022, doi: 10.1093/llc/fqab079.
[5] B. Herwijayanti, D. E. Ratnawati, and L. Muflikhah, “Online news classification using TF-IDF weighting and cosine similarity,”
(Indonesian) “Klasifikasi berita online dengan menggunakan pembobotan TF-IDF dan cosine similarity,” Jurnal Pengembangan
Teknologi Informasi dan Ilmu Komputer, vol. 2, no. 1, pp. 306–312, 2018, [Online]. Available: http://j-ptiik.ub.ac.id.
[6] T. Hariguna, B. Fandy, S. Utomo, J. Pol, S. Watumas, and P. J. Pol, “Implementation of information retrieval Indonesian text
documents using the vector space model,” 2012, [Online]. Available: https://www.researchgate.net/publication/304605781.
[7] S. Salmon, D. Paseru, and V. Kumenap, “Implementasi metode vector space model pada search engine perpustakaan,” Prosiding
SISFOTEK, vol. 4, no. 1, pp. 84–92, 2020, [Online]. Available: http://seminar.iaii.or.id/index.php/SISFOTEK/article/view/158.
[8] D. B. Lomet, L. Gravano, A. Halevy, and G. Weikum, “Editorial board (Associate Editors),” Journal of the Franklin Institute,
vol. 358, no. 4, p. i, 2021, doi: 10.1016/s0016-0032(21)00098-3.
[9] O. Shahmirzadi, A. Lugowski, and K. Younge, “Text similarity in vector space models: A comparative study,” Proceedings - 18th
IEEE International Conference on Machine Learning and Applications, ICMLA 2019, pp. 659–666, 2019, doi:
10.1109/ICMLA.2019.00120.
[10] R. Singh and S. Singh, “Text similarity measures in news articles by vector space model using NLP,” Journal of The Institution of
Engineers (India): Series B, vol. 102, no. 2, pp. 329–338, 2021, doi: 10.1007/s40031-020-00501-5.
[11] S. Jaiswal and R. Kumar, Learning Django web development : from idea to prototype, a learner’s guide for web development wih
the Django application framework. 2015.
[12] K. M. Vamsi, P. Lokesh, K. N. Reddy, and P. Swetha, “Visualization of real-world enterprise data using Python Django
framework,” IOP Conference Series: Materials Science and Engineering, vol. 1042, no. 1, p. 012019, 2021, doi: 10.1088/1757-
899x/1042/1/012019.
[13] A. N. Rahimah, D. S. Rusdianto, and M. T. Ananta, "Development of a web-based reading room management system using the
Django framework (Case study: reading room, Faculty of Computer Science, Brawijaya University)" (Indonesian)
“Pengembangan sistem pengelolaan ruang baca berbasis web dengan menggunakan Django framework (Studi kasus: ruang baca
fakultas ilmu komputer Universitas Brawijaya),” Jurnal Pengembangan Teknologi Informasi Dan Ilmu Komputer, vol. 2, no. 5,
pp 4439-4446, 2019.
[14] I. Taufik, M. Jaenudin, F. U. Badriyah, B. Subaeki, and O. T. Kurahman, “The search for science and technology verses in qur’an
and hadith,” Bulletin of Electrical Engineering and Informatics, vol. 10, no. 2, pp. 1008–1014, 2021, doi:
10.11591/eei.v10i2.2629.
[15] V. Basmalah Wicaksono, R. Saptono, and S. Widya Sihwi, “Comparative analysis of vector space model and weighted tree
similarity methods with cosine similarity in the case of searching for basic treatment guideline information at community health
centers" (Indonesian) "Analisis perbandingan metode vector space model dan weighted tree similarity dengan cosine similarity
pada kasus pencarian informasi pedoman pengobatan dasar di puskesmas,” Jurnal Teknologi & Informasi ITSmart, vol. 4, no. 2,
p. 73, 2016, doi: 10.20961/its.v4i2.1768.
[16] A. Sonita and M. Sari, ““Implementation of sequential searching algorithm for searching letter numbers in electronic archive
systems" (Indonesian) "Implementasi algoritma sequential searching untuk pencarian nomor surat pada sistem arsip elektronik,”
Pseudocode, vol. 5, no. 1, pp. 1–9, 2018, doi: 10.33369/pseudocode.5.1.1-9.

Vector space model, term frequency-inverse document frequency with linear search…(Ichsan Taufik)
314  ISSN: 2722-3221

[17] L. Heryawan, D. Novitaningrum, K. R. Nastiti, and S. N. Mahmudah, “Medical record document search with TF-IDF and vector
space model (VSM),” International Journal on Advanced Science, Engineering and Information Technology, vol. 14, no. 3, pp.
847–852, 2024, doi: 10.18517/ijaseit.14.3.19606.
[18] H. H. Batubara, "Utilization of the encyclopedia of hadith books of the 9 imams as a medium and source for learning hadith,"
(Indonesian) “Pemanfaatan ensiklopedi hadis kitab 9 imam sebagai media dan sumber belajar hadis,” Muallimuna : Jurnal
Madrasah Ibtidaiyah, vol. 2, no. 2, p. 63, 2017, doi: 10.31602/muallimuna.v2i2.769.
[19] Anna and A. Hendini, “Implementation of vector space model in karaoke machine search system,” (Indonesian) “Implementasi
vector space model pada sistem pencarian mesin karaoke,” Evolusi : Jurnal Sains dan Manajemen, vol. 6, no. 1, 2018, doi:
10.31294/evolusi.v6i1.3535.
[20] Bania Amburika, Y. H. Chrisnanto, and W. Uriawan, "Vector space model (VSM) technique in determining the handling of the
impact of online games on children," (Indonesian) “Teknik vector space model (VSM) dalam penentuan penanganan dampak
game online pada anak,” Prosiding SNST ke-7 Tahun 2016, vol. 1, no. 1, pp. 10–27, 2016, [Online]. Available:
http://cogsys.imm.dtu.dk/thor/projects/multimedia/textmining/node5.html.
[21] H. Hu et al., “Horizontal and vertical crossover of sine cosine algorithm with quick moves for optimization and feature selection,”
Journal of Computational Design and Engineering, vol. 9, no. 6, pp. 2524–2555, 2022, doi: 10.1093/jcde/qwac119.
[22] S. Saha, M. K.Bhaumik, and S. Das, “A new modified linear search algorithm,” International Journal of Mathematics Trends and
Technology, vol. 65, no. 12, pp. 148–152, 2019, doi: 10.14445/22315373/ijmtt-v65i12p516.
[23] L. Goyal, K. Sharma, T. Rai, B. Gupta, and N. Arora, “Left-right mid way linear searching algorithm,” International Journal of
Computer Applications, vol. 177, no. 45, pp. 22–26, 2020, doi: 10.5120/ijca2020919939.
[24] A. Muhazir, M. Fakhriza, and E. Sutejo, "Implementation of sequential method in searching for goods distribution in cargo
integration system," (Indonesian) “Implementasi metode sequential dalam pencarian pendistribusian barang pada cargo
integration sistem,” Jurnal dan Penelitian Teknik Informatika, vol. 2, no. 2, pp. 24–30, 2017.
[25] O. Lukyanchikov, E. Pluzhnik, S. Payain, and E. Nikulchev, “Using object-relational mapping to create the distributed databases
in a hybrid cloud infrastructure,” International Journal of Advanced Computer Science and Applications, vol. 5, no. 12, 2014, doi:
10.14569/ijacsa.2014.051208.
[26] A. Joshi and S. Kukreti, “Object relational mapping in comparison to traditional data access techniques,” International Journal of
Scientific & Engineering Research, vol. 5, no. 6, pp. 540–543, 2014, [Online]. Available: http://www.ijser.org.
[27] K. M. Ting, “Confusion matrix,” Encyclopedia of Machine Learning, pp. 209–209, 2011, doi: 10.1007/978-0-387-30164-8_157.

BIOGRAPHIES OF AUTHORS

Ichsan Taufik is a lecturer in Department Informatics at UIN Sunan Gunung

Djati Bandung. Ichsan has graduated from Informatics Department at IAIN Sunan Gunung
Djati. He received his M.T. Degree from Institut Teknologi Bandung. Research areas:
Information System, Algorithm and Programming, Software Engineering. He can be
contacted at email: ichsan@uinsgd.ac.id.

Agra received his first degree from UIN Sunan Gunung Djati Bandung. His
academic qualification is S.T. His working at consultant IT in Jakarta. He can be contacted at
email: agra0108@gmail.com.

Yana Aditia Gerhana received his first degree from Sekolah Tinggi Teknologi
Garut, Informatic, Garut, in 2003. He has also Master degree from STMIK Likmi,
Information System, Bandung, in 2009. The Dr. degree from the Vocation and Technology
Education in Universitas Pendidikan Indonesia (UPI), Indonesia in 2016. He is currently a
computer science lecuturer. His main research interests focus on Information System,
Multimedia, Vocational and Tehcnology Educataion, Information Retrieval Data Mining, and
Text Mining. He can be contacted at email: yanagerhana@uinsgd.ac.id.

Comput Sci Inf Technol, Vol. 5, No. 3, November 2024: 306-314