Garuda - Garba Rujukan Digital

Enhancing Predictive Models: An In-depth Analysis of Feature Selection Techniques Coupled with Boosting Algorithms Neny Sulistianingsih; Galih Hendro Martono
MATRIK : Jurnal Manajemen, Teknik Informatika dan Rekayasa Komputer Vol. 23 No. 2 (2024)
Publisher : LPPM Universitas Bumigora

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.30812/matrik.v23i2.3788

This research addresses the critical need to enhance predictive models for fetal health classification using Cardiotocography (CTG) data. The literature review underscores challenges in imbalanced labels, feature selection, and efficient data handling. This paper aims to enhance predictive models for fetal health classification using Cardiotocography (CTG) data by addressing challenges related to imbalanced labels, feature selection, and efficient data handling. The study uses Recursive Feature Elimination (RFE) and boosting algorithms (XGBoost, AdaBoost, LightGBM, CATBoost, and Histogram-Based Boosting) to refine model performance. The results reveal notable variations in precision, Recall, F1-Score, accuracy, and AUC across different algorithms and RFE applications. Notably, Random Forest with XGBoost exhibits superior performance in precision (0.940), Recall (0.890), F1-Score (0.920), accuracy (0.950), and AUC (0.960). Conversely, Logistic Regression with AdaBoost demonstrates lower performance. The absence of RFE also impacts model effectiveness. In conclusion, the study successfully employs RFE and boosting algorithms to enhance fetal health classification models, contributing valuable insights for improved prenatal diagnosis.

Optimizing Water Hyacinth as Organic Fertilizer to Support Zero Waste and Green Economy Initiatives Martono, Galih Hendro; Neny Sulistianingsih; Ni Putu Sinta Dewi
ABDIMAS: Jurnal Pengabdian Masyarakat Vol. 8 No. 2 (2025): ABDIMAS UMTAS: Jurnal Pengabdian Kepada Masyarakat
Publisher : LPPM Universitas Muhammadiyah Tasikmalaya

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.35568/abdimas.v8i2.6510

The overgrowth of water hyacinth (Eichhornia crassipes) in Batujai Dam, located in West Lombok Regency, has become a serious environmental concern. Its uncontrolled spread has disrupted water flow, limited irrigation functions, and negatively impacted aquatic biodiversity. However, instead of treating it as a problem, this community service activity focused on turning the weed into a helpful resource—specifically, a raw material for producing organic fertilizer. Purpose: The aim of this community service was to raise public awareness and provide training on managing water hyacinth sustainably while creating added value for the local economy. The program was conducted with a small-scale fertilizer producer in Central Lombok. Method: Using a Participatory Action Research (PAR) approach, the activity involved the local community at every stage—from identifying the issue, designing solutions, and implementing the processing techniques to evaluating the results together. This approach was chosen to build community ownership and ensure the continuity of the efforts after the program ended. Result: As part of the process, around 1,000 kilograms of water hyacinth were harvested, sun-dried, chopped, and composted using Trichoderma spp. After fermentation, the community produced 20 liters of liquid fertilizer and 400 kilograms of solid compost. Conclusion: Beyond its environmental impact, the activity opened up opportunities for alternative income and promoted the concept of zero waste. It also encouraged the community to see local ecological issues not as obstacles but as opportunities to support green and sustainable living.

Sentiment Analysis of Service and Facility Satisfaction at Computer Lab of Universitas Bumigora Using Indobert Mundika, Eko; Martono, Galih Hendro; Rismayati, Ria
Journal of Artificial Intelligence and Software Engineering Vol 5, No 2 (2025): June
Publisher : Politeknik Negeri Lhokseumawe

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.30811/jaise.v5i2.6798

Computer laboratories have a strategic role in supporting the technology-based learning process at Bumigora University. To understand the extent to which the available services and facilities meet students' expectations, this study conducted a sentiment analysis of student reviews using the IndoBERT model, an artificial intelligence-based Natural Language Processing (NLP) approach. Data was obtained from a questionnaire focusing on aspects of laboratory services and facilities, then analyzed to classify opinions into positive, negative, and neutral sentiments. The analysis results show the dominance of positive sentiments, indicating that computer laboratories have generally met student expectations, especially in supporting practicum activities. The IndoBERT model used was able to achieve 85% accuracy, demonstrating its effectiveness in reliably identifying opinion trends. These findings provide a comprehensive picture of student perceptions, and serve as an important basis for managers in formulating strategies to improve the quality of laboratory services and facilities so that a conducive and relevant learning experience can be maintained.

Analysis of the Effectiveness of Traditional and Ensemble Machine Learning Models for Mushroom Classification Sulistianingsih, Neny; Martono, Galih Hendro
J-INTECH ( Journal of Information and Technology) Vol 13 No 01 (2025): J-Intech : Journal of Information and Technology
Publisher : LPPM STIKI MALANG

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.32664/j-intech.v13i01.1851

The classification of edible versus poisonous mushrooms presents a critical challenge in the domains of applied biology and public health, particularly due to the serious implications of misidentification. This research employs the UCI Mushroom Dataset to evaluate and compare the effectiveness of several machine learning models, including traditional algorithms like Logistic Regression, Decision Tree, Random Forest, Support Vector Machine, K-Nearest Neighbors and Naïve Bayes, as well as advanced ensemble techniques such as Stacking and Voting Classifier. Notably, both Random Forest and Stacking achieved flawless accuracy, reaching 100%, underscoring the high predictive capacity of these models in complex categorical scenarios. Conversely, Naïve Bayes exhibited significantly weaker performance—achieving only 59.8% accuracy—likely due to its underlying assumption of feature independence, which does not hold for this dataset. The ensemble learning approaches, including the combination of Stacking and Bagging, not only preserved but also enhanced model robustness and generalization. These methods effectively leverage the complementary strengths of individual learners to yield more accurate and stable predictions while mitigating overfitting risks. Comparative analysis with previous research confirms the consistency of these findings and reinforces the viability of ensemble strategies for handling intricate classification tasks. Overall, this study highlights the importance of algorithm selection tailored to data characteristics and supports the use of ensemble learning to boost predictive reliability.

Optimizing Autism Spectrum Disorder Identification with Dimensionality Reduction Technique and K-Medoid Martono, Galih Hendro; Sulistianingsih, Neny
JURNAL INFOTEL Vol 16 No 4 (2024): November 2024
Publisher : LPPM INSTITUT TEKNOLOGI TELKOM PURWOKERTO

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.20895/infotel.v16i4.1142

This research addresses the challenges of diagnosing and treating Autism Spectrum Disorder (ASD) using dimensionality reduction techniques and machine learning approaches. Challenges in social interaction, communication, and repetitive behaviours characterize ASD. The dimension reduction used in this research aims to identify what features influence autism cases. Several dimension data reduction techniques used in this research include PCA, Isomap, t-SNE, LLE, and factor analysis, using metrics such as Purity, silhouette score, and the Fowlkes-Mallows index. The machine learning approach applied in this study is k-medoid. By employing this method, our goal is to pinpoint the unique characteristics of autism that may facilitate the detection and diagnosis process. The data used in this research is a dataset collected for autism screening in adults. This dataset contains 20 features: ten behavioural features (AQ-10-Adult) and ten individual characteristics. The results indicate that Factor Analysis outperforms other methods based on purity metrics. However, due to data structure issues, the t-SNE method cannot be evaluated using purity metrics. PCA and LLE consistently provide stable silhouette scores across different values. The Fowlkes-Mallows index results closely align, but t-SNE tends to yield lower values. The choice of algorithm requires careful consideration of preferred metrics and data characteristics. Factor analysis is adequate for Purity, while PCA and LLE consistently perform well. This research aims to improve the accuracy of ASD identification, thereby enhancing diagnostic and treatment precision.

Pelatihan Penggunaan Kuesioner Usability untuk Pengukuran Kualitas Perangkat Lunak bagi Mahasiswa dan Masyarakat Umum: Training on the Use of Usability Questionnaire for Measuring Software Quality for Students and the General Public Wahyuningrum, Tenia; Martono, Galih Hendro; Wardhana, Helna
PengabdianMu: Jurnal Ilmiah Pengabdian kepada Masyarakat Vol. 10 No. 8 (2025): PengabdianMu: Jurnal Ilmiah Pengabdian kepada Masyarakat
Publisher : Institute for Research and Community Services Universitas Muhammadiyah Palangkaraya

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.33084/pengabdianmu.v10i8.9598

Agile development ensures quality software is on time and according to user expectations. Therefore, software testing is essential before launch, including usability testing as part of the user experience. However, the many choices of questionnaires in usability testing make it difficult for developers to determine the appropriate method. This community service is a training activity that aims to improve the understanding of students and the general public, especially those involved in software development related to usability testing. The training is conducted face-to-face with counseling, namely interactive lectures, discussions, case studies, and quizzes. The training materials include an introduction to software quality and usability evaluation, respondent selection techniques, an introduction to types of usability questionnaires, and exercises in calculating and interpreting usability scores using the Usability Metric for User Experience (UMUX-Lite) questionnaire. The training results showed that the level of understanding was still low (31.07%); this result was due to several obstacles, namely the imbalance in the number of participants and facilitators, the relatively short discussion time, and not all participants being active in the activities. The strategy for future improvements is to conduct remedial sessions, re-evaluate learning methods and media, and divide participants into small groups with adequate training facilitators.

Perbandingan Algoritma Sarima dan Prophet Untuk Peramalan Trend Penjualan Voucher Game Online Rizki, M; Priyanto, Dadang; Martono, Galih Hendro; Sulistianingsih, Neny; Syahrir, Moch
Jurnal Minfo Polgan Vol. 14 No. 2 (2025): Artikel Penelitian
Publisher : Politeknik Ganesha Medan

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.33395/jmp.v14i2.15083

Industri game online terus mengalami perkembangan pesat, mendorong kebutuhan akan sistem peramalan yang akurat untuk mendukung pengambilan keputusan strategis dalam manajemen penjualan dan promosi. Studi ini bertujuan untuk membandingkan kinerja dua algoritma peramalan deret waktu, yaitu Seasonal Autoregressive Integrated Moving Average (SARIMA) dan Prophet, dalam memprediksi tren penjualan voucher game online di platform Kiyystore. Data yang digunakan dalam penelitian ini mencakup transaksi historis dari tahun 2022 hingga 2024, dengan total 5,530 data penjualan. Studi ini menerapkan metodologi Cross Industry Standard Process for Data Mining (CRISP DM) yang terdiri dari tahap pemahaman bisnis, pemrosesan data, pemodelan, dan evaluasi. Model SARIMA dipilih karena kemampuannya untuk menangkap pola musiman dan tren dalam data stasioner. Sementara itu, Prophet digunakan karena dirancang untuk menangani tren non-linear, pola musiman, dan anomali secara otomatis. Evaluasi kinerja dari kedua algoritma dilakukan menggunakan dua metrik utama, yaitu Mean Absolute Error (MAE) dan Root Mean Squared Error (RMSE). Hasil penelitian menunjukkan bahwa Prophet unggul dalam metrik MAE dengan nilai 0,7054, yang menunjukkan kinerja yang lebih baik dalam meminimalkan kesalahan rata-rata. Di sisi lain, SARIMA menunjukkan keunggulan dalam metrik RMSE dengan nilai 0,9514, yang berarti model ini lebih efektif dalam menangani kesalahan besar atau pencilan dalam prediksi. Studi ini memberikan kontribusi penting dalam pemilihan metode peramalan yang sesuai dengan karakteristik data. Dengan memahami keunggulan masing-masing algoritma, pelaku industri game online dapat lebih optimal dalam merencanakan strategi stok dan promosi, sehingga meningkatkan efisiensi dan daya saing bisnis secara keseluruhan

Penerapan ARAS dan TOPSIS pada Sistem Pendukung Keputusan Untuk Seleksi Penerimaan Anggota PPK di KPU Sumbawa Barat Ramadhan, Rahmat Adi Mulya; Husain, Husain; Martono, Galih Hendro
RIGGS: Journal of Artificial Intelligence and Digital Business Vol. 4 No. 3 (2025): Agustus - October
Publisher : Prodi Bisnis Digital Universitas Pahlawan Tuanku Tambusai

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.31004/riggs.v4i3.3015

Dalam rangka meningkatkan transparansi dan akuntabilitas pada proses seleksi Panitia Pemilihan Kecamatan (PPK), Komisi Pemilihan Umum (KPU) Kabupaten Sumbawa Barat membutuhkan sistem seleksi yang objektif, terukur, dan dapat dipertanggungjawabkan. Selama ini seleksi masih dilakukan secara konvensional sehingga berpotensi menimbulkan subjektivitas dan kurang efisien. Penelitian ini bertujuan membangun sistem pendukung keputusan berbasis metode Additive Ratio Assessment (ARAS) dan Technique for Order Preference by Similarity to Ideal Solution (TOPSIS) guna membantu proses pemilihan anggota PPK secara sistematis dan adil. Metode penelitian menggunakan pendekatan rekayasa perangkat lunak dengan model pengembangan Waterfall, yang meliputi analisis kebutuhan, perancangan, implementasi, pengujian, dan evaluasi. Data kriteria seleksi diperoleh melalui observasi, wawancara, dan studi pustaka. Sistem dibangun menggunakan PHP dan MySQL, dengan ARAS dan TOPSIS sebagai dasar perhitungan multi-kriteria, mencakup indikator seperti pengalaman organisasi, kemampuan manajerial, integritas, pemahaman kepemiluan, serta keterampilan komunikasi.Hasil penelitian menunjukkan sistem mampu menghasilkan peringkat alternatif calon anggota PPK secara objektif. Uji kepuasan pengguna memperoleh skor 85,33% dengan kategori “sangat setuju”, menandakan penerimaan positif dari aspek fungsionalitas, kemudahan penggunaan, dan akurasi penilaian. Kesimpulannya, penerapan ARAS dan TOPSIS terbukti saling melengkapi dalam pemeringkatan calon PPK. Sistem ini menjadikan seleksi lebih transparan, efisien, dan kredibel, sekaligus berpotensi diterapkan pada rekrutmen sumber daya manusia di instansi pemerintahan lainnya.

Analisis Dampak Pelatihan Canva dalam Komunikasi Visual Sulistianingsih, Neny; Hasbullah, Hasbullah; Martono, Galih Hendro
Jurnal Pengabdian Pada Masyarakat IPTEKS Vol. 1 No. 2: Jurnal Pengabdian Pada Masyarakat IPTEKS, Juni 2024
Publisher : CV. Global Cendekia Inti

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.71094/jppmi.v1i2.52

The use of Canva in educational communication has garnered attention, yet research exploring its use in announcements and communication with students remains limited. This study aims to optimize visual communication by providing Canva usage training to academic and program staff, with a focus on announcements and student communication. The engagement method follows a participatory approach and Service learning. Questionnaire results show a significant increase in confidence levels and graphic design abilities post-training. Positive social and behavioral changes are also observed. From a theoretical perspective, these findings are supported by visual design theories and service learning. Conclusions indicate that Canva training is effective in enhancing the quality of visual communication between educational institutions and students. Recommendations include continuing and expanding training and monitoring implementation outcomes.

Enhancing Predictive Models: An In-depth Analysis of Feature Selection Techniques Coupled with Boosting Algorithms Sulistianingsih, Neny; Martono, Galih Hendro
MATRIK : Jurnal Manajemen, Teknik Informatika dan Rekayasa Komputer Vol. 23 No. 2 (2024)
Publisher : Universitas Bumigora

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.30812/matrik.v23i2.3788

This research addresses the critical need to enhance predictive models for fetal health classification using Cardiotocography (CTG) data. The literature review underscores challenges in imbalanced labels, feature selection, and efficient data handling. This paper aims to enhance predictive models for fetal health classification using Cardiotocography (CTG) data by addressing challenges related to imbalanced labels, feature selection, and efficient data handling. The study uses Recursive Feature Elimination (RFE) and boosting algorithms (XGBoost, AdaBoost, LightGBM, CATBoost, and Histogram-Based Boosting) to refine model performance. The results reveal notable variations in precision, Recall, F1-Score, accuracy, and AUC across different algorithms and RFE applications. Notably, Random Forest with XGBoost exhibits superior performance in precision (0.940), Recall (0.890), F1-Score (0.920), accuracy (0.950), and AUC (0.960). Conversely, Logistic Regression with AdaBoost demonstrates lower performance. The absence of RFE also impacts model effectiveness. In conclusion, the study successfully employs RFE and boosting algorithms to enhance fetal health classification models, contributing valuable insights for improved prenatal diagnosis.

Title

Found 25 Documents
Search

Abstract

Abstract

Abstract

Abstract

Abstract

Abstract

Abstract

Abstract

Abstract

Abstract

Title Search

Found 25 Documents Search

Abstract

Abstract

Abstract

Abstract

Abstract

Abstract

Abstract

Abstract

Abstract

Abstract

Title

Found 25 Documents
Search