Gede Aditra Pradnyana
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Research Interests
Gede Aditra Pradnyana has a total of 56 co-authors. Here are the top co-authors by number of shared publications:
- Gede Saindra Santyadiputra
Universitas Pendidikan Ganesha
5 shared publications - I. Made Agus Wirawan
Universitas Pendidikan Ganesha
5 shared publications - I. Gede Mahendra Darmawiguna
Universitas Pendidikan Ganesha
4 shared publications I Gede Mahendra Darmawiguna
Universitas Pendidikan Ganesha
3 shared publications- I Gede Dedy Prasetia
Universitas Pendidikan Ganesha
2 shared publications - Ign Edo Paska
Universitas Pendidikan Ganesha
2 shared publications
Latest Publications
- Open Access
- 1 August 2026
- Published by Universitas Muhammadiyah Malang in Kinetik: Game Technology, Information System, Computer Network, Computing, Electronics, and Control
Indonesian Pre-trained Language Models with OCEAN-aware Query Expansion Ranking for Personality Prediction from Social Media Text
Personality prediction from social media text has become a critical topic in computational social science, as online posts frequently reflect individual behavioral and psychological characteristics. However, predicting personality from Indonesian social media material remains challenging due to informal language, slang, abbreviations, noisy expressions, and implicit personality-related cues. This study proposed a new OCEAN personality prediction framework that integrates pre-trained language models with an OCEAN-aware Query Expansion Ranking feature learning mechanism. Unlike conventional approaches that rely mainly on contextual embeddings, the proposed framework introduced trait-specific class-discriminative lexical features to strengthen personality-related representation. The prediction task was structured as five separate binary classification problems, with each personality attribute divided into High and Low groups.. The experiments were conducted in two stages. First, numerous pre-trained language models, including IndoBERT, IndoBERTweet, Indonesian RoBERTa, and multilingual BERT, were fine-tuned and compared to identify the most suitable contextual representation model. The best baseline model, multilingual BERT, achieved an average accuracy of 75.48% and an average F1-score of 73.58%. Second, multilingual BERT was integrated with class-discriminative lexical features generated by the proposed OCEAN-aware Query Expansion Ranking mechanism. The trait-specific configuration improved the average accuracy to 78.76% and the average F1-score to 76.48%. These results demonstrated that integrating contextual semantic representations with OCEAN-aware lexical feature learning enhanced personality prediction from Indonesian social media material while also offering a more clear representation of trait-relevant language signals.
- 1 August 2026
- Published by Universitas Muhammadiyah Malang in Kinetik: Game Technology, Information System, Computer Network, Computing, Electronics, and Control
Open AccessIndonesian Pre-trained Language Models with OCEAN-aware Query Expansion Ranking for Personality Prediction from Social Media Text
Personality prediction from social media text has become a critical topic in computational social science, as online posts frequently reflect individual behavioral and psychological characteristics. However, predicting personality from Indonesian social media material remains challenging due to informal language, slang, abbreviations, noisy expressions, and implicit personality-related cues. This study proposed a new OCEAN personality prediction framework that integrates pre-trained language models with an OCEAN-aware Query Expansion Ranking feature learning mechanism. Unlike conventional approaches that rely mainly on contextual embeddings, the proposed framework introduced trait-specific class-discriminative lexical features to strengthen personality-related representation. The prediction task was structured as five separate binary classification problems, with each personality attribute divided into High and Low groups.. The experiments were conducted in two stages. First, numerous pre-trained language models, including IndoBERT, IndoBERTweet, Indonesian RoBERTa, and multilingual BERT, were fine-tuned and compared to identify the most suitable contextual representation model. The best baseline model, multilingual BERT, achieved an average accuracy of 75.48% and an average F1-score of 73.58%. Second, multilingual BERT was integrated with class-discriminative lexical features generated by the proposed OCEAN-aware Query Expansion Ranking mechanism. The trait-specific configuration improved the average accuracy to 78.76% and the average F1-score to 76.48%. These results demonstrated that integrating contextual semantic representations with OCEAN-aware lexical feature learning enhanced personality prediction from Indonesian social media material while also offering a more clear representation of trait-relevant language signals.
- Open Access
- 28 July 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Access
- Vol. 13, 133465-133482
Revealing Depression Through Social Media via Adaptive Gated Cross-Modal Fusion Augmented With Insights From Personality Traits
The escalating prevalence of mental health disorders, particularly depression, underscores the urgent need for early and accurate detection systems. The widespread use of social media has produced an extensive repository of user-generated content, offering critical insight into individuals’ emotional and psychological states. This development enables novel approaches for automated depression detection. However, existing multimodal depression detection approaches often adopt rigid fusion strategies and disregard individual differences in expressive behavior by adopting generalized, one-size-fits-all frameworks. To bridge this gap, we introduce DeXMAG, a novel personalized depression detection framework that integrates a Cross-Modal Attention mechanism with an Adaptive Gated Fusion strategy. DeXMAG effectively combines textual content, visual imagery, and user-level personality traits to capture the heterogeneous and individualized patterns of depressive expression across modalities. Leveraging domain-specific transformer-based language models and a convolutional neural networks-based model, we extract modality-specific features and fuse them via a novel adaptive gating mechanism. Unlike traditional fusion techniques, our model leverages modality-aware attention signals and gating coefficients to enable flexible and personalized integration of features. An extensive ablation study reveals that the most substantial performance gain occurs when DeXMAG is augmented with insights from Myers–Briggs Type Indicator (MBTI) personality traits in conjunction with textual and visual features. Comprehensive experiments on multimodal social media datasets demonstrate that DeXMAG consistently outperforms unimodal baselines and state-of-the-art approaches, achieving an accuracy of 95.20% and an F1-score of 95.94%. These results reinforce the effectiveness of adaptive gated fusion and personality-aware modeling in advancing robust and personalized depression detection.
- 28 July 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Access
- Vol. 13, 133465-133482
Open AccessRevealing Depression Through Social Media via Adaptive Gated Cross-Modal Fusion Augmented With Insights From Personality Traits
The escalating prevalence of mental health disorders, particularly depression, underscores the urgent need for early and accurate detection systems. The widespread use of social media has produced an extensive repository of user-generated content, offering critical insight into individuals’ emotional and psychological states. This development enables novel approaches for automated depression detection. However, existing multimodal depression detection approaches often adopt rigid fusion strategies and disregard individual differences in expressive behavior by adopting generalized, one-size-fits-all frameworks. To bridge this gap, we introduce DeXMAG, a novel personalized depression detection framework that integrates a Cross-Modal Attention mechanism with an Adaptive Gated Fusion strategy. DeXMAG effectively combines textual content, visual imagery, and user-level personality traits to capture the heterogeneous and individualized patterns of depressive expression across modalities. Leveraging domain-specific transformer-based language models and a convolutional neural networks-based model, we extract modality-specific features and fuse them via a novel adaptive gating mechanism. Unlike traditional fusion techniques, our model leverages modality-aware attention signals and gating coefficients to enable flexible and personalized integration of features. An extensive ablation study reveals that the most substantial performance gain occurs when DeXMAG is augmented with insights from Myers–Briggs Type Indicator (MBTI) personality traits in conjunction with textual and visual features. Comprehensive experiments on multimodal social media datasets demonstrate that DeXMAG consistently outperforms unimodal baselines and state-of-the-art approaches, achieving an accuracy of 95.20% and an F1-score of 95.94%. These results reinforce the effectiveness of adaptive gated fusion and personality-aware modeling in advancing robust and personalized depression detection.
- Open Access
- 1 June 2024
- Published by Ikatan Ahli Informatika Indonesia (IAII) in Jurnal RESTI (Rekayasa Sistem dan Teknologi Informasi)
- Vol. 8 (3) , 321-332
A Middleware Applications Design for Health Information Sharing
The interoperability between electronic health records (EHR) and electronic medical records (EMR) from various healthcare facilities for comprehensive patient care is important. However, integrating such systems, including the need for interoperability standards, data privacy, and security, is a highly challenging task, especially since patient rights in data access must be considered. The primary problem addressed is the challenge of integrating electronic health records (EHR) and electronic medical records (EMR) within various healthcare facilities to ensure comprehensive patient care while maintaining data privacy, security, and adherence to patient rights. This work presents an innovative application for consolidating patient health records from various medical facilities. Facilitates seamless data access, improving the efficiency of healthcare care delivery. The GGD approach was used in developing the prototype to ensure that the delivered product was able to meet the user requirements. Four phases are divided into six stages used in this method: research, modeling, requirements definition, framework definition, refinement, and support. The evaluation involved two phases, back-end and front-end testing, using white-box and black-box testing. White-box testing delivers an average frame-rendering rate of up to 56 fps, and black-box testing has shown 100% successful results in the given task. In conclusion, the Med-OID prototype was successfully developed. Integrates and securely transmits medical records across various healthcare services, demonstrating significant potential to enhance personalized medicine and healthcare coordination. The evaluations underscored the robustness of the prototype and its ability to improve interoperability and data sharing in healthcare systems.
- 1 June 2024
- Published by Ikatan Ahli Informatika Indonesia (IAII) in Jurnal RESTI (Rekayasa Sistem dan Teknologi Informasi)
- Vol. 8 (3) , 321-332
Open AccessA Middleware Applications Design for Health Information Sharing
The interoperability between electronic health records (EHR) and electronic medical records (EMR) from various healthcare facilities for comprehensive patient care is important. However, integrating such systems, including the need for interoperability standards, data privacy, and security, is a highly challenging task, especially since patient rights in data access must be considered. The primary problem addressed is the challenge of integrating electronic health records (EHR) and electronic medical records (EMR) within various healthcare facilities to ensure comprehensive patient care while maintaining data privacy, security, and adherence to patient rights. This work presents an innovative application for consolidating patient health records from various medical facilities. Facilitates seamless data access, improving the efficiency of healthcare care delivery. The GGD approach was used in developing the prototype to ensure that the delivered product was able to meet the user requirements. Four phases are divided into six stages used in this method: research, modeling, requirements definition, framework definition, refinement, and support. The evaluation involved two phases, back-end and front-end testing, using white-box and black-box testing. White-box testing delivers an average frame-rendering rate of up to 56 fps, and black-box testing has shown 100% successful results in the given task. In conclusion, the Med-OID prototype was successfully developed. Integrates and securely transmits medical records across various healthcare services, demonstrating significant potential to enhance personalized medicine and healthcare coordination. The evaluations underscored the robustness of the prototype and its ability to improve interoperability and data sharing in healthcare systems.
- Open Access
- 30 April 2023
- Published by Institut Bisnis dan Teknologi Indonesia in SINTECH (Science and Information Technology) Journal
- Vol. 6 (1) , 58-67
Optimasi Parameter Support Vector Machine Dengan Algoritma Genetika Untuk Analisis Sentimen Pada Media Sosial Instagram
Social media is an online media that users use to interact with each other by expressing themselves by giving comments, and one example is Instagram. All the collected comments will form a public opinion. This opinion can be used with sentiment analysis to become information. The method commonly used to carry out sentiment analysis is classification using machine learning. One of the machine learning that is often used is the Support Vector Machine (SVM). However, on non-linear problems such as sentiment analysis, SVM requires the kernel to map vectors into high-dimensional spaces to solve non-linear problems. The problem faced in using the kernel is to choose the optimal parameters for the classification model to produce a good classification model. This study proposes a new approach to obtain optimal parameters for SVM using Genetic Algorithm (GA). This study designed an SVM-GA classification model from the data collection, processing, classification, and evaluation stages. The results showed that the best accuracy produced with parameters optimized with the genetic algorithm was 81.6%, or an increase of 2.4% from the SVM sentiment analysis method without GA optimization.
- 30 April 2023
- Published by Institut Bisnis dan Teknologi Indonesia in SINTECH (Science and Information Technology) Journal
- Vol. 6 (1) , 58-67
Open AccessOptimasi Parameter Support Vector Machine Dengan Algoritma Genetika Untuk Analisis Sentimen Pada Media Sosial Instagram
Social media is an online media that users use to interact with each other by expressing themselves by giving comments, and one example is Instagram. All the collected comments will form a public opinion. This opinion can be used with sentiment analysis to become information. The method commonly used to carry out sentiment analysis is classification using machine learning. One of the machine learning that is often used is the Support Vector Machine (SVM). However, on non-linear problems such as sentiment analysis, SVM requires the kernel to map vectors into high-dimensional spaces to solve non-linear problems. The problem faced in using the kernel is to choose the optimal parameters for the classification model to produce a good classification model. This study proposes a new approach to obtain optimal parameters for SVM using Genetic Algorithm (GA). This study designed an SVM-GA classification model from the data collection, processing, classification, and evaluation stages. The results showed that the best accuracy produced with parameters optimized with the genetic algorithm was 81.6%, or an increase of 2.4% from the SVM sentiment analysis method without GA optimization.
- 21 December 2021
- Published by Atlantis Press SARL
- p. 316-320
Web-Based System for Bali Tourism Sentiment Analysis during The Covid-19 Pandemic using Django Web Framework and Naive Bayes Method
- 21 December 2021
- Published by Atlantis Press SARL
- p. 316-320
Web-Based System for Bali Tourism Sentiment Analysis during The Covid-19 Pandemic using Django Web Framework and Naive Bayes Method
- 21 December 2021
- Published by Atlantis Press SARL
- p. 116-121
The Development of an Android-Based Kata Kolok Sign Language Dictionary
- 21 December 2021
- Published by Atlantis Press SARL
- p. 116-121
The Development of an Android-Based Kata Kolok Sign Language Dictionary
- Open Access
- 30 September 2021
- Published by Universitas Udayana in Jurnal Ilmu Komputer
- Vol. 14 (2) , 88-98
Kombinasi Metode Naive Bayes dan K-Medoid dalam Memprediksi Penjurusan Siswa di Sekolah Menengah Atas
Penjurusan merupakan suatu proses penempatan atau penyaluran dalam pemilihan program pengajaran kepada siswa. Tujuan dari penentuan penjurusan itu sendiri adalah agar kelak dikemudian hari pelajaran yang diberikan kepada siswa lebih terarah. SMA Laboratorium Undiksha memiliki permasalahan dalam penentuan penjurusan dan pembagian kelas siswa. Proses penentuan jurusan membutuhkan waktu yang cukup lama, masih menggunakan perhitungan manual di excel dalam menentukan penjurusan siswa.. Permasalahan tersebut dapat diatasi dengan Pengembangan Sistem Prediksi Penjurusan Kelas Siswa Menggunakan Kombinasi Metode Naive Bayes Dan K-Medoid. Ada 22 kriteria yang digunakan dalam penentuan jurusan kelas siswa yaitu, jenis kelamin, nilai raport semester 3 sampai dengan semester 5, nilai matematika semester 3 sampai dengan semester 5, nilai ipa semester 3sampai dengan semester, nilai ips semester 3 sampai dengan semester 5, nilai bahasa indonesia semester 3 sampai dengan semester 5, nilai bahasa inggris semester 3 sampai dengan semester 5, minat siswa 1, minat siswa 2 dan minat orang tua. Berdasarkan hasil perhitungan kinerja metode, didapatkan nilai akurasi sebesar 76% yang menunjukkan bahwa metode yang digunakan memiliki nilai akurasi yang cukup baik dalam memprediksi jurusan siswa. Precision yang dihasilkan sebesar 84,33% menunjukkan kategori data yang diklasifikasi telah sesuai dengan kategori yang sebenarnya. Recall yang dihasilkan sebesar 70,67% menunjukkan tingkat keberhasilan metode dalam mengenali suatu kategori sudah baik.
- 30 September 2021
- Published by Universitas Udayana in Jurnal Ilmu Komputer
- Vol. 14 (2) , 88-98
Open AccessKombinasi Metode Naive Bayes dan K-Medoid dalam Memprediksi Penjurusan Siswa di Sekolah Menengah Atas
Penjurusan merupakan suatu proses penempatan atau penyaluran dalam pemilihan program pengajaran kepada siswa. Tujuan dari penentuan penjurusan itu sendiri adalah agar kelak dikemudian hari pelajaran yang diberikan kepada siswa lebih terarah. SMA Laboratorium Undiksha memiliki permasalahan dalam penentuan penjurusan dan pembagian kelas siswa. Proses penentuan jurusan membutuhkan waktu yang cukup lama, masih menggunakan perhitungan manual di excel dalam menentukan penjurusan siswa.. Permasalahan tersebut dapat diatasi dengan Pengembangan Sistem Prediksi Penjurusan Kelas Siswa Menggunakan Kombinasi Metode Naive Bayes Dan K-Medoid. Ada 22 kriteria yang digunakan dalam penentuan jurusan kelas siswa yaitu, jenis kelamin, nilai raport semester 3 sampai dengan semester 5, nilai matematika semester 3 sampai dengan semester 5, nilai ipa semester 3sampai dengan semester, nilai ips semester 3 sampai dengan semester 5, nilai bahasa indonesia semester 3 sampai dengan semester 5, nilai bahasa inggris semester 3 sampai dengan semester 5, minat siswa 1, minat siswa 2 dan minat orang tua. Berdasarkan hasil perhitungan kinerja metode, didapatkan nilai akurasi sebesar 76% yang menunjukkan bahwa metode yang digunakan memiliki nilai akurasi yang cukup baik dalam memprediksi jurusan siswa. Precision yang dihasilkan sebesar 84,33% menunjukkan kategori data yang diklasifikasi telah sesuai dengan kategori yang sebenarnya. Recall yang dihasilkan sebesar 70,67% menunjukkan tingkat keberhasilan metode dalam mengenali suatu kategori sudah baik.
- 18 September 2021
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 280-284
Towards Healthcare Data Sharing: An e-Health Integration Effort in Indonesian District
The purpose of this work was to provide an integration scheme of healthcare systems, connecting the public health centres to the general hospital and provide the required data aggregation to entitled government institutions. A middle-ware-based system has been developed using a rapid application development model and tested using several scenarios. The results have shown that the developed system could deliver the expected process while the newly provided executive dashboard system provided insight into the processed data. The developed web services are ready to be used in further data exchanges, but several steps must be ascertained beforehand. Problems, challenges, and opportunities regarding the future of data sharing among institutions in Indonesia were also discussed. The services that have been developed in this work shows a promising implementation of an integrated system, especially in healthcare, as part of the implementation of the broader e-Gov-ernance enactment. The emphasis on the role of patients in the built system becomes an essential key for further development.
- 18 September 2021
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 280-284
Towards Healthcare Data Sharing: An e-Health Integration Effort in Indonesian District
The purpose of this work was to provide an integration scheme of healthcare systems, connecting the public health centres to the general hospital and provide the required data aggregation to entitled government institutions. A middle-ware-based system has been developed using a rapid application development model and tested using several scenarios. The results have shown that the developed system could deliver the expected process while the newly provided executive dashboard system provided insight into the processed data. The developed web services are ready to be used in further data exchanges, but several steps must be ascertained beforehand. Problems, challenges, and opportunities regarding the future of data sharing among institutions in Indonesia were also discussed. The services that have been developed in this work shows a promising implementation of an integrated system, especially in healthcare, as part of the implementation of the broader e-Gov-ernance enactment. The emphasis on the role of patients in the built system becomes an essential key for further development.
- Open Access
- 27 July 2021
- Published by Universitas Pendidikan Ganesha in International Journal of Natural Science and Engineering
- Vol. 5 (2) , 77-86
Prototype of Water Quality Monitoring for Grouper Fish Pond Based on Microcontroller Arduino
Prototype monitoring of grouper pond water quality is intended to improve the water quality of grouper ponds if there is a change in air quality such as acidity, air temperature, and air turbidity level, which can be caused by fish or fish that can be used on fish breathing fish that can make fish become stressful and cannot multiply properly and can cause death in fish. To overcome these problems, a technology that can be studied and controlled according to the standards is needed. It can also facilitate pond farmers to facilitate pond air quality and the automation of standardization of water quality indicators. This development uses internet-related matters in data communication between prototypes with farm farmers. The development model used is prototyping. For the testing process, (3) trials are conducted: (1) evaluation testing, with test results, prototypes, get a percentage of 100% with a very effective category. (2) validity testing, with a margin of error of 4.38% and the overall sensor test results under 4.38%. (3) practically test, obtained by the results of 92.2% of respondents stated that monitoring of this prototype is very practical.
- 27 July 2021
- Published by Universitas Pendidikan Ganesha in International Journal of Natural Science and Engineering
- Vol. 5 (2) , 77-86
Open AccessPrototype of Water Quality Monitoring for Grouper Fish Pond Based on Microcontroller Arduino
Prototype monitoring of grouper pond water quality is intended to improve the water quality of grouper ponds if there is a change in air quality such as acidity, air temperature, and air turbidity level, which can be caused by fish or fish that can be used on fish breathing fish that can make fish become stressful and cannot multiply properly and can cause death in fish. To overcome these problems, a technology that can be studied and controlled according to the standards is needed. It can also facilitate pond farmers to facilitate pond air quality and the automation of standardization of water quality indicators. This development uses internet-related matters in data communication between prototypes with farm farmers. The development model used is prototyping. For the testing process, (3) trials are conducted: (1) evaluation testing, with test results, prototypes, get a percentage of 100% with a very effective category. (2) validity testing, with a margin of error of 4.38% and the overall sensor test results under 4.38%. (3) practically test, obtained by the results of 92.2% of respondents stated that monitoring of this prototype is very practical.
- Open Access
- 1 March 2021
- Published by IOP Publishing in Journal of Physics: Conference Series
- Vol. 1810 (1) , 012033
Performance analysis of support vector machines with polynomial kernel for sentiment polarity identification: A case study in lecturer’s performance questionnaire
The lecturers’ performance evaluation process can be carried out using an open questionnaire filled out by students at the end of the semester. In this questionnaire, students can provide an assessment in the form of comments, suggestions, and criticism of the ’lecturer’s performance, which can describe the level of student satisfaction with the lecture process. Conducting assessments or analyses on the open questionnaire entries manually will certainly impact the high costs time and energy. Sentiment polarity identification is a process in sentiment analysis that classifies text into a sentence or document and then determines whether the opinion expressed is positive, negative or neutral. In this research, a sentiment polarity detection system was developed in a lecturer evaluation questionnaire using the Support Vector Machine (SVM) method with a polynomial kernel. The test results showed that the SVM method’s performance with the Polynomial kernel was strongly influenced by the value of the learning rate parameter, the maximum iteration, and the degree, with the optimal parameter values, respectively, 0.001, 200, and 0.3. The use of optimal parameter values in the process of identifying sentiment polarity obtained an accuracy value of 84.88%.
- 1 March 2021
- Published by IOP Publishing in Journal of Physics: Conference Series
- Vol. 1810 (1) , 012033
Open AccessPerformance analysis of support vector machines with polynomial kernel for sentiment polarity identification: A case study in lecturer’s performance questionnaire
The lecturers’ performance evaluation process can be carried out using an open questionnaire filled out by students at the end of the semester. In this questionnaire, students can provide an assessment in the form of comments, suggestions, and criticism of the ’lecturer’s performance, which can describe the level of student satisfaction with the lecture process. Conducting assessments or analyses on the open questionnaire entries manually will certainly impact the high costs time and energy. Sentiment polarity identification is a process in sentiment analysis that classifies text into a sentence or document and then determines whether the opinion expressed is positive, negative or neutral. In this research, a sentiment polarity detection system was developed in a lecturer evaluation questionnaire using the Support Vector Machine (SVM) method with a polynomial kernel. The test results showed that the SVM method’s performance with the Polynomial kernel was strongly influenced by the value of the learning rate parameter, the maximum iteration, and the degree, with the optimal parameter values, respectively, 0.001, 200, and 0.3. The use of optimal parameter values in the process of identifying sentiment polarity obtained an accuracy value of 84.88%.