Implementation of Knowledge Graph as a Representation of Public Sentiment Analysis Toward AI-Generated Art

Finna Nur Nandia; I Kadek Dwi Nuryana

doi:10.31004/riggs.v5i2.11349

Authors

Finna Nur Nandia Universitas Negeri Surabaya
I Kadek Dwi Nuryana Universitas Negeri Surabaya

DOI:

https://doi.org/10.31004/riggs.v5i2.11349

Keywords:

AI-Generated Art, Sentiment Analysis, IndoBERT, LDA, Knowledge Graph, SNA

Abstract

The rapid advancement of generative AI technology has elicited diverse public responses on social media, particularly toward AI-generated art, which has significantly impacted Indonesia's creative industry. This study aims to analysed the sentiment of Indonesian society on platform X toward AI-generated art and to represent the interconnections among sentiment entities through a Knowledge Graph (KG). The approach integrates three primary methods within the Knowledge Discovery in Databases (KDD) framework: sentiment classification using the IndoBERT model, topic modelling using Latent Dirichlet Allocation (LDA), and KG construction based on Social Network Analysis (SNA). The dataset consists of 4,765 Indonesian-language tweets that underwent pre-processing. Sentiment analysis results indicate a dominance of negative sentiment (45.4%) over positive sentiment (42.5%), with the IndoBERT model achieving 69% accuracy on a three-class classification task. Topic modelling produced 18 distinctive topics (9 negative, 9 positive), validated through Two-Stage Similarity validation. Negative topics are dominated by issues of economic impact on illustrators, copyright infringement, and artistic style theft, while positive topics reflect appreciation for AI as a creative tool. The constructed KG comprises 144 nodes and 329 edges with a modularity score of 0.4556, reflecting 7 meaningful thematic communities. SNA evaluation reveals that 'ilustrasi' (illustration) is the most central entity (degree centrality = 12.621), while negative issues dominate the central positions among topic nodes. This study demonstrates that the integration of IndoBERT, LDA, and KG is capable of uncovering hidden relational patterns in public opinion that cannot be obtained through conventional sentiment analysis alone.

Downloads

Download data is not yet available.

References

[1] N. Muhamad, “Indonesia, Penyumbang Kunjungan Aplikasi AI Terbanyak ke-3 di Dunia,” Databoks Katadata, Jan. 31, 2024. [Online]. Available: https://databoks.katadata.co.id/teknologi-telekomunikasi/statistik/a49ed3eb121983b/indonesia-penyumbang-kunjungan-aplikasi-ai-terbanyak-ke-3-di-dunia

[2] C. Ramadhan, V. Atina, and H. Permatasari, “Analisis Perbandingan Model CNN dan IndoBERT Dalam Sentimen Berita Politik Indonesia,” dalam Prosiding Seminar Nasional Teknologi Informasi dan Bisnis (SENATIB) 2025, Surakarta, Indonesia, Jul. 2025, doi: 10.47701/v1r9ka69.

[3] D. Alfatah, “Application Of Transformer Model For Sentiment Detection On Indonesian Twitter Data,” Jurnal Komputer, vol. 2, no. 2, 2024, doi: 10.70963/jk.v2i2.99.

[4] S. S. Nabila, “Analisis Sentimen Publik terhadap Gambar Berbasis AI di Twitter Menggunakan Metode Naive Bayes,” preprint, 2025. [Online]. Available: https://www.researchgate.net/publication/391848610_Analisis_Sentimen_Publik_terhadap_Gambar_Berbasis_AI_di_Twitter_Menggunakan_Metode_Naive_Bayes/citation/download.

[5] Y. Sahria and D. H. Fudholi, “Pemodelan Pengetahuan Graph Database Untuk Jejaring Penelitian Kesehatan di Indonesia,” Jurnal Media Informatika Budidarma, vol. 4, no. 3, pp. 604–611, Jul. 2020, doi: 10.30865/mib.v4i3.2183.

[6] D. Prismala and I. K. D. Nuryana, “Analisis Opinion Mining pada Topik ChatGPT di Aplikasi X dengan Pendekatan Algoritma SVM Berbasis Lexicon,” Journal of Informatics and Computer Science (JINACS), vol. 6, no. 2, pp. 479–492, Jul. 2024.

[7] U. Khairani, V. Mutiawani, and H. Ahmadian, “Pengaruh Tahapan Preprocessing Terhadap Model IndoBERT dan IndoBERTweet untuk Mendeteksi Emosi pada Komentar Akun Berita Instagram,” Jurnal Teknologi Informasi dan Ilmu Komputer, vol. 11, no. 4, pp. 887–894, Aug. 2024, doi: 10.25126/jtiik.1148315.

[8] F. Koto, A. Rahimi, J. H. Lau, and T. Baldwin, "IndoLEM and IndoBERT: A benchmark dataset and pre-trained language model for Indonesian NLP," in Proc. 1st Conf. Asia-Pacific Chapter Assoc. Comput. Linguist. 10th Int. Joint Conf. Natural Language Process. (AACL-IJCNLP), Suzhou, China, Dec. 2020, pp. 843–857.

[9] W. X. Zhao, J. Jiang, J. Weng, J. He, E.-P. Lim, H. Yan, and X. Li, "Comparing Twitter and traditional media using topic models," in Proc. 33rd Eur. Conf. Inf. Retrieval (ECIR 2011), Lecture Notes in Computer Science, vol. 6611, Dublin, Ireland, Apr. 2011, pp. 338–349.

[10] M. Á. Serrano, M. Boguñá, and A. Vespignani, "Extracting the multiscale backbone of complex weighted networks," Proc. Natl. Acad. Sci. USA, vol. 106, no. 16, pp. 6483–6488, Apr. 2009.

[11] N. Mamatha, S. S. Sunitha, and M. D. Shivakumar, "Graph theory and its role in social network analysis," World J. Adv. Res. Rev., vol. 21, no. 2, pp. 2088–2093, 2024.

[12] V. E. Sidauruk and W. Herowati, "IndoBERT-based sentiment analysis of political discourse on platform X: The case of Prabowo-Gibran administration," J. Appl. Inform. Comput. (JAIC), vol. 10, no. 1, pp. 673–683, 2024.

[13] B. Christopher, Analisis Sentimen Penggunaan AI-Generated Artwork di Twitter Menggunakan IndoBERT. Tangerang: Universitas Multimedia Nusantara, 2025.

[14] S. J. Blair, Y. Bi, and M. D. Mulvenna, "Aggregated topic models for increasing social media topic coherence," Appl. Intell., vol. 50, no. 1, pp. 138–156, Jan. 2020.

[15] A. M. Sadri, S. Hasan, S. V. Ukkusuri, and J. E. S. Lopez, “Analyzing Social Interaction Networks from Twitter for Planned Special Events,” arXiv preprint arXiv:1704.02489, Apr. 2017.

[16] J. Xie and B. K. Szymanski, "Community detection using a neighborhood strength driven label propagation algorithm," in Proc. 2011 IEEE Netw. Sci. Workshop (NSW), New York, NY, USA, Jun. 2011, pp. 188–195.

[17] J. Vizcarra, K. Kozaki, M. Torres-Ruiz, and R. Quintero, “Knowledge-Based Sentiment Analysis and Visualization on Social Networks,” New Generation Computing, vol. 38, no. 4, pp. 701–742, Oct. 2020.