Article Per Year (5 Year)
p-Index From 2021 - 2026
2.679
P-Index
This Author published in this journals
All Journal Sainteks Jurnal Informatika dan Teknik Elektro Terapan Sistemasi: Jurnal Sistem Informasi Jurnal Nasional Komputasi dan Teknologi Informasi Progresif: Jurnal Ilmiah Komputer Journal of Innovation and Future Technology (IFTECH) Jurnal Inovasi Penelitian dan Pengabdian Masyarakat Urecol Journal. Part E: Engineering Jurnal Abdimas Indonesia Jurnal Bisnis Digital dan Sistem Informasi Jurnal Ilmu Komputer dan Matematika
Taftazani Ghazi Pratama
Universitas Muhammadiyah Kudus
Religion Aerospace Engineering Humanities Automotive Engineering Chemical Engineering, Chemistry & Bioengineering Chemistry Civil Engineering, Building, Construction & Architecture Computer Science & IT Control & Systems Engineering Decision Sciences, Operations Research & Management Economics, Econometrics & Finance Education Electrical & Electronics Engineering Engineering Environmental Science Health Professions Industrial & Manufacturing Engineering Languange, Linguistic, Communication & Media Law, Crime, Criminology & Criminal Justice Library & Information Science Mathematics Mechanical Engineering Medicine & Pharmacology Neuroscience Nursing Public Health Social Sciences Transportation Other

Published : 14 Documents Claim Missing Document
Claim Missing Document
Check
Articles

Found 14 Documents
Search

 1   2 
DETEKSI DINI ASD(AUTISM SPECTRUM DISORDER) MENGGUNAKAN MACHINE LEARNING Taftazani Ghazi Pratama; Achmad Ridwan; Agung Prihandono
JURNAL ILMU KOMPUTER DAN MATEMATIKA Vol 4, No 2 (2023): JURNAL ILMU KOMPUTER DAN MATEMATIKA
Publisher : Universitas Muhammadiyah Kudus

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.26751/jikoma.v4i2.1998

Abstract

Deteksi dini ASD pada seorang balita sangat membantu orang tua untuk mengetahui kembang tumbuh anak. Pada penelitian ini bertujuan untuk  deteksi dini ASD  menggunakan Naive Bayes dan KNN yang diterapkan pada dataset Autism screening data for toddlers.  Penelitian ini dilakukan melalui tahapan pra pengolahan, pembagian data training 80% dan testing 20%, pembuatan model, dan evaluasi dari model yang dibuat.  Hasil evaluasi model yang dibuat  menunjukkan bahwa KNN memperoleh nilai akurasi, sensitivitas, dan spesifisitas lebih tinggi daripada Naive Bayes. KNN memperoleh nilai akurasi sebesar  95,73%, sensitivitas sebesar 93,84%, dan  spesifisitas100%. Hal ini mengindikasikan bahwa KNN dapat membantu dalam deteksi dini pada seorang balita dengan kinerja yang baik. 
Analisis Perbandingan Kinerja Algoritma Naïve Bayes Dan KNN Untuk Memprediksi Penyakit Diabetes Osama Maulana Haq; Achmad Ridwan; Taftazani Ghazi Pratama
Progresif: Jurnal Ilmiah Komputer Vol 21, No 1 (2025): Februari
Publisher : STMIK Banjarbaru

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.35889/progresif.v21i1.2424

Abstract

Diabetes is a chronic disease affecting various age groups with a risk of fatal complications. Accurate diagnosis is a crucial initial step in management; however, the gradual progression of symptoms often leads to delayed detection. This study compares the accuracy of the Naïve Bayes and K-Nearest Neighbors (KNN) algorithms in predicting diabetes using a dataset from Kaggle. Naïve Bayes was chosen for its ability to handle large datasets, missing values, irrelevant attributes, and noise, while KNN offers simplicity in implementation. The results show that KNN achieves a higher accuracy of 79% compared to Naïve Bayes at 76%. Therefore, KNN is recommended for diabetes prediction based on this dataset.Keywords: Diabetes; Naïve Bayes, K-Nearest Neighbors; Accuracy AbstrakDiabetes merupakan penyakit kronis yang menyerang berbagai usia dengan risiko komplikasi fatal. Diagnosis yang akurat menjadi langkah awal penting untuk pengelolaan, namun gejala yang berkembang perlahan sering menyebabkan keterlambatan deteksi. Penelitian ini membandingkan akurasi algoritma Naïve Bayes dan K-Nearest Neighbors (KNN) dalam memprediksi diabetes menggunakan dataset dari Kaggle. Naïve Bayes dipilih karena kemampuannya menangani dataset besar, data hilang, atribut tidak relevan, dan noise, sedangkan KNN menawarkan kesederhanaan implementasi. Hasil pengujian menunjukkan bahwa KNN memiliki akurasi lebih tinggi sebesar 79% dibandingkan Naïve Bayes yang mencapai 76%. Dengan demikian, KNN lebih direkomendasikan untuk prediksi diabetes berdasarkan dataset ini.Kata Kunci: Diabetes; Naïve Bayes; K-Nearest Neighbors; Akurasi
Klasifikasi Gangguan Tidur Menggunakan Algoritma XGBoost dengan SMOTE dan Grid Search Moh. Indra Kholid Khoirusshofi; Fida Maisa Hana; Taftazani Ghazi Pratama
Sainteks Vol. 23 No. 1 (2026): April
Publisher : Lembaga Penelitian dan Pengabdian Masyarakat (LPPM)

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.30595/sainteks.v23i1.30171

Abstract

Gangguan tidur merupakan permasalahan kesehatan yang memiliki prevalensi tinggi dan berdampak signifikan terhadap kualitas hidup, sehingga diperlukan sistem deteksi dini yang akurat dan efisien. Perkembangan machine learning membuka peluang untuk membangun model klasifikasi gangguan tidur berbasis data, namun ketidakseimbangan kelas pada dataset medis sering menjadi tantangan yang menurunkan performa model, khususnya pada kelas minoritas. Penelitian ini bertujuan membangun model klasifikasi gangguan tidur menggunakan algoritma Extreme Gradient Boosting (XGBoost) dengan penerapan Synthetic Minority Oversampling Technique (SMOTE) dan optimasi hiperparameter menggunakan Grid Search. Dataset yang digunakan adalah Sleep Health and Lifestyle Dataset yang terdiri dari 374 data dengan tiga kelas target, yaitu Insomnia, No Disorder, dan Sleep Apnea. Penelitian ini menguji empat skenario model, yaitu XGBoost tanpa SMOTE dan Grid Search, XGBoost dengan SMOTE, XGBoost dengan Grid Search, serta kombinasi SMOTE dan Grid Search. Evaluasi kinerja model dilakukan menggunakan metrik accuracy, precision, recall, F1-score, Confusion Matrix, serta ROC Curve. Hasil penelitian menunjukkan bahwa penerapan SMOTE meningkatkan sensitivitas model terhadap kelas minoritas, sedangkan optimasi hiperparameter menggunakan Grid Search meningkatkan stabilitas dan akurasi model secara keseluruhan. Kombinasi SMOTE dan Grid Search menghasilkan performa terbaik dengan akurasi mencapai 97% serta nilai precision, recall, dan F1-score yang seimbang pada seluruh kelas. Selain itu, evaluasi menggunakan ROC Curve menunjukkan nilai AUC pada rentang 0,99 hingga 1,00, yang mengindikasikan kemampuan model yang sangat baik dalam membedakan setiap kelas. Hasil ini menunjukkan bahwa pendekatan yang diusulkan mampu meningkatkan performa klasifikasi gangguan tidur dan berpotensi menjadi alternatif dalam pengembangan sistem deteksi dini gangguan tidur.
Detecting Muslim Students Mental Health with an Islamic Educational Approach using Machine Learning Taftazani Ghazi Pratama; Toni Ardi Rafsanjani; Riana Putri Rahmawati; Helmi Imaduddin
SISTEMASI Vol 15, No 1 (2026): Sistemasi: Jurnal Sistem Informasi
Publisher : Universitas Islam Indragiri

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.32520/stmsi.v15i1.5732

Abstract

Mental health among university students has become a major concern in higher education, particularly in the post-pandemic era, which has left students facing various academic, social, and psychological pressures. Unfortunately, efforts for early detection of mental health issues on campus remain limited, especially in the context of Muslim students who live within an Islamic cultural framework. This study offers an innovative approach by integrating advanced machine learning technology with the depth of Islamic educational values to develop an early detection system that is not only accurate but also humanistic and contextually relevant. The dataset for this study was obtained through a survey of 127 students at Universitas Muhammadiyah Kudus, including variables related to psychological conditions and the intensity of religious practices, used to detect whether students experience mental health problems or maintain good mental health. The research methodology includes data collection, preprocessing, feature analysis, model development using classification algorithms such as Random Forest, SVM, KNN, and Decision Tree, model performance optimization using GridSearchCV, and evaluation. Evaluation of the four models indicated that prior to optimization, SVM and KNN achieved the best performance, both with an accuracy of 88.46%. After optimization with GridSearchCV, SVM became the top-performing model, achieving an accuracy improvement of more than 5%, reaching 94.05%. Feature analysis revealed that levels of anxiety, fatigue, and religious practices such as prayer and dhikr were the primary determinants in mapping students’ mental health conditions. These findings suggest that Islamic values such as tawakkul (trust in God), sabr (patience), and syukur (gratitude) are not merely theological concepts but can also serve as scientific instruments, converted into predictive features in data-driven technologies. This study demonstrates that an SVM model optimized with GridSearchCV is effective in detecting university students’ mental health and has the potential to serve as an early warning system in Islamic campus settings.
Co-Authors Achmad Ridwan Achmad Ridwan, Achmad Agung Prihandono Andi Kurniawan Anwari . Ashri Maulida Rahmawati Dian Rosita Edi Wibowo Suwandi Fida Maisa Hana Hafni Aulida Helmi Imaduddin Hermawan Abdillah Hamka Mahdarul Huda Ahmad Moh. Indra Kholid Khoirusshofi Muhamad Jauhar Osama Maulana Haq Purbowati Purbowati Riana Putri Rahmawati Sam Saputra, Vernanda Saputra, Vernanda Sam Soma Setiawan Ponco Nugroho Syahrul Azmi Nawaf Toni Ardi Rafsanjani Widya Cholid Wahyudin