Topic Modeling for Support Ticket using Latent Dirichlet Allocation

Wiranto, Wiranto; Uswatunnisa, Mila Rosyida

doi:10.29207/resti.v6i6.4542

Cited by 4 publications

(6 citation statements)

References 23 publications

(27 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Such multifaceted examinations of technical, social, and educational lenses are prudent when assessing complex innovations such as ChatGPT. Tracking conceptual shifts through dynamic topic modeling will illuminate whether current skepticism gives way to embracing transformational opportunities as research coalesces around long-term, ethical integrations (Wiranto & Uswatunnisa, 2022;Yi et al, 2020).…”

Section: Discussionmentioning

confidence: 99%

Charting the Growth and Structure of Early ChatGPT-Education Research: A Bibliometric Study

Watrianthos,

Triono Ahmad,

Muskhir

2023

JITE:IIP

View full text Add to dashboard Cite

Aim/Purpose: The purpose of this article is to provide an overview and analysis of the emerging research landscape surrounding the integration of ChatGPT into education. The main problem appears to be that this is a new, rapidly developing research area for which there is no comprehensive synthesis of the current literature. The aim of the article is to fill this gap by conducting a timely bibliometric study to map publication trends, influential works, themes, and opportunities, thus representing the growth and structure of ChatGPT educational research. Background: This article addresses the issue of the lack of a comprehensive synthesis of the new research on ChatGPT in education by conducting a bibliometric analysis. Specifically, the authors use statistical and network analysis techniques to examine the patterns of publication, citation, and keywords and map the growth, contributions, themes, structure, and opportunities in this evolving field. The bibliometric approach provides a comprehensive, evidence-based overview of the current state of the literature to uncover trends and gaps and help researchers improve their understanding of appropriate and effective applications of ChatGPT in educational contexts. Methodology: The authors used bibliometric analysis as the primary method to summarize the new research on ChatGPT in education. We searched the database of the Web of Science Core Collection to find 51 relevant documents from 2023 that included ChatGPT in the title and were classified as ‘educational research.’ The sample consisted of these 51 documents, including articles, early access articles, editorials, reviews, and letters. Statistical techniques examined publication, citation, and keyword patterns. Network analysis visualized citation and co-occurrence networks to reveal intellectual structure. The multifaceted bibliometric approach allowed a comprehensive study of the sample from a productive, conceptual, and intellectual perspective. Contribution: This article conducts comprehensive bibliometric analysis of this emerging research area and synthesizes publication, citation, and keyword data to map the growth and structure of the literature. The results reveal important trends, such as the rapid growth of publications since the release of ChatGPT, initial authorship patterns, the focus on higher education applications, and distinct research clusters around pedagogical, ethical, and assessment issues. Visualizing citation networks identifies seminal studies while mapping co-occurrence clarifies conceptual relationships between topics. The comparative analysis highlights the differences between document types, topics, and time periods. Knowledge mapping highlights gaps in the literature, such as lack of focus on K-12 contexts, and highlights opportunities for further research. Findings: Key findings from this bibliometric analysis of the emerging research land-scape surrounding ChatGPT integration in education include the following: • Since ChatGPT was released in late 2022, the number of releases has increased significantly, indicating rapid growth in this emerging space. • The most cited authors initially came primarily from Anthropic, but over time, the citations spread throughout the research community. • The topics focused primarily on higher education applications, with a clear focus on pedagogical strategies, ethical risks, and implications for assessment. • Citation networks visualized seminal studies, while the co-occurrence of keywords clarified conceptual connections. • Gaps such as applications in the K-12 context were uncovered, and opportunities for further research were highlighted. • The literature is rapidly evolving and requires ongoing monitoring of the development of this field. In general, the analysis presents the productivity, contributors, themes, struc-ture, and opportunities in this emerging area around the integration of ChatGPT in education based on current scientific evidence. The key findings focus on the growing early interest, gaps and developments that can provide insight for researchers and educators. Recommendations for Practitioners: Practitioners should carefully integrate ChatGPT into education based on new evidence, carefully assess contextual applicability, and proactively develop guidelines for ethical and equitable implementation. Ongoing advice, impact monitoring, and research partnerships are crucial to informing best practices. Educators must be vigilant for risks such as privacy, student well-being, and competence impairment while staying abreast of advances in knowledge to dynamically adapt integration strategies. The introduction should empower diverse learners through measured, integrative approaches based on continuous contextual analysis and ethical principles. Recommendation for Researchers: This article recommends that researchers conduct more studies in under-researched contexts, use multiple methods to capture nuanced impacts, increase focus on responsible integration strategies, develop tailored assessments, conduct interdisciplinary collaborations, monitor long-term adoption, mix with interactive explain and publish open access technologies, help guide adoption pathways through actionable studies, and synthesize the exponentially growing literature through updated systematic reviews. Impact on Society: The rapid publication growth and prevailing optimism suggest that the integration of ChatGPT into education will accelerate, increasing the need for rigorous research that guides ethical, responsible innovations that avoid risks and improve outcomes in all educational contexts. The findings have broader implications for guiding adoption trajectories through ongoing evidence synthesis and expanded investigations in under-researched areas to address knowledge gaps. Ultimately, continued monitoring and updated guidance are critical to ensure that ChatGPT’s educational penetration progresses carefully by maximizing benefits and minimizing harms in rapidly evolving AI-powered learning ecosystems. Future Research: Based on the basic mapping provided by this paper, recommended research directions include longitudinal impact studies, research tailored to under-researched contexts such as K-12, qualitative research to capture stakeholder perspectives, development and testing of AI-calibrated assessments as well as explorations that combine conversational and interactive learning technologies, updated systematic reviews, and co-designed implementation research that explain pedagogical strategies that ethically unlock learning potential while mitigating risks in diverse educational environments. Such multilayered tracking can provide critical insights to guide context-specific, responsible ChatGPT integration and monitor impact within rapidly evolving AI-powered education ecosystems.

show abstract

Section: Discussionmentioning

confidence: 99%

Charting the Growth and Structure of Early ChatGPT-Education Research: A Bibliometric Study

Watrianthos,

Triono Ahmad,

Muskhir

2023

JITE:IIP

View full text Add to dashboard Cite

show abstract

“…Nilai coherence dapat mengukur hubungan antar kata dalam pemodelan topik, evaluasi topik terbaik dinilai berdasarkan nilai coherence tertinggi yang diperoleh [7]. Di samping itu, penelitian ini juga melakukan evaluasi menggunakan pendekatan human judgement untuk menginterpretasikan distribusi kata dari hasil pemodelan topik [17]. Dalam konteks penilaian manusia, penelitian ini menggunakan pendekatan intrusi kata untuk mengevaluasi topik yang dinilai memiliki makna sama dan saling berkaitan satu sama lain sehingga dapat diidentifikasi oleh manusia [21].…”

Section: Topic Modeling Menggunakan Latent Dirichlet Allocationunclassified

“…Hal ini berarti bahwa dalam satu topik terdapat 3 interpretasi yang berbeda. Dalam penilaian manusia, kondisi seperti ini dapat dikatakan buruk sehingga tidak dapat diinterpretasikan [17].…”

Section: Topikunclassified

“…Jumlah topik terbaik terdapat pada topik 3 sehingga menghasilkan topik seperti covidindonesia, covid_19, pandemic di Indonesia, dan pembahasan mutasi virus covid-19. Penelitian lainnya [17] menggunakan LDA untuk menguji apakah LDA dapat diterapkan pada tiket dukungan. Analisis menggunakan 5 parameter secara acak yaitu number of topic, nilai alpha, nilai beta, iterasi, dan random_seed untuk mendapatkan nilai coherence tertinggi.…”

unclassified

See 1 more Smart Citation

Pemodelan Topik pada Media Berita Online Menggunakan Latent Dirichlet Allocation (Studi Kasus Merek Somethinc)

Puspita,

Shiddieq,

Roji

2024

MALCOM

View full text Add to dashboard Cite

Somethinc merupakan salah satu merek kosmetik lokal di Indonesia yang aktif memanfaatkan media, seperti berita online untuk menyampaikan informasi terkini seputar merek. Dari banyaknya berita online mengenai merek Somethinc, sering kali topik dan tren yang sedang dibahas tidak menggambarkan informasi secara keseluruhan. Untuk menganalisis topik yang paling sering dibahas dalam berita online mengenai merek Somethinc, peneliti menggunakan metode topic modeling, yaitu Latent Dirichlet Allocation, yang dinilai lebih unggul dalam menghasilkan topik secara terstruktur. Penelitian ini memanfaatkan nilai coherence untuk menganalisis dan mengevaluasi jumlah topik terbaik, selanjutnya pendekatan human judgement digunakan untuk menginterpretasikan topik. Hasil analisis kemudian divisualisasikan secara interaktif menggunakan pyLDAvis, untuk mengetahui persebaran kata dari setiap topik. Berdasarkan hasil penelitian, jumlah topik terbaik terdapat pada topik 6 dengan nilai coherence sebesar 0.404. Keenam topik tersebut diinterpretasikan berdasarkan pendekatan human judgement, menghasilkan topik-topik meliputi produk skincare untuk kulit berjerawat, penghargaan brand kecantikan terbaik, kolaborasi produk, produk perawatan kulit dan kecantikan, kampanye pemasaran produk, dan brand lokal dengan produk perawatan kecantikan. Dapat disimpulkan bahwa jumlah topik 6 menghasilkan topik-topik yang relevan mengenai merek Somethinc.

show abstract

“…That is where the IT help desk software comes to the picture. IT helpdesk ticketing software is being used by companies to provide a centralized facility to troubleshoot and facilitate solutions to IT-related issues [6].…”

Section: Introduction: -mentioning

confidence: 99%

Help Desk Support Ticket and Issue Management

Lakshya Sharma, Sharik Ahmad

2023

IJRITCC

View full text Add to dashboard Cite

IT companies require methodical approaches to handle the growing volume of customer-reported software issues and service requests. An enormous backlog of unresolved issues has the potential to drastically raise software development and maintenance expenses. An IT company therefore need a clearly defined customer support model. It acts as a link between the client and the business from beginning to end. The support industry also used this to track and record any procedures and solutions used to close tickets. With the Helpdesk Ticketing System, you can use an online platform to solve questions in a web-based environment. To ensure seamless and efficient operation, the user can submit tickets for even the smallest questions. Both within the company and remotely, the ticketing tool is usable. which can be accessed by anyone in organization or end user.

show abstract

Topic Modeling for Support Ticket using Latent Dirichlet Allocation

Cited by 4 publications

References 23 publications

Charting the Growth and Structure of Early ChatGPT-Education Research: A Bibliometric Study

Charting the Growth and Structure of Early ChatGPT-Education Research: A Bibliometric Study

Pemodelan Topik pada Media Berita Online Menggunakan Latent Dirichlet Allocation (Studi Kasus Merek Somethinc)

Help Desk Support Ticket and Issue Management

Contact Info

Product

Resources

About