Claim Missing Document
Check
Articles

Found 37 Documents
Search

Model Prediksi Jumlah Penjualan Pelumas Mesin Di PT. X Dengan Algoritma Naïve Bayes Purnama, Nilam; Fitri Insani; Elin Haerani; Iis Afrianty
Computer Science and Information Technology Vol 5 No 3 (2024): Jurnal Computer Science and Information Technology (CoSciTech)
Publisher : Universitas Muhammadiyah Riau

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.37859/coscitech.v5i3.8250

Abstract

Machine lubricants are essential materials used to reduce friction between two moving surfaces, improve machine efficiency, and extend the lifespan of components. This study aims to predict the sales volume of machine lubricants at PT. X using the Naïve Bayes algorithm. The data used includes attributes such as year, month, material description, total allocation, realization, and remaining allocation, with a total of 3,006 data points obtained from PT. X's Warehouse Management System (WMS). The model was tested using the 10-Fold Cross Validation method and testsing without such validation. The test results show an accuracy of 71% with 10-Fold Cross Validation, compared to 14% without validation. Additional testing showed an accuracy of 5%, with RMSE of 124.71 and MAPE of 0.95. Based on these results, it is recommended to optimize data preprocessing, such as handling data imbalance and feature normalization, to improve prediction accuracy. Furthermore, using more diverse validation techniques, such as stratified cross-validation, can provide more stable evaluations. Given that predictions are influenced solely by historical data, it is recommended to periodically update the data to keep the model relevant and accurate. This research is expected to assist PT. X in planning sales strategies and managing lubricant stock more effectively.
Implementasi Algoritma Improve Apriori Terhadap Keluarga Beresiko Stunting Muhammad Habib Nazlis; Fitri Insani; Alwis Nazir; Iis Afrianty
Computer Science and Information Technology Vol 5 No 3 (2024): Jurnal Computer Science and Information Technology (CoSciTech)
Publisher : Universitas Muhammadiyah Riau

Show Abstract | Download Original | Original Source | Check in Google Scholar

Abstract

Stunting is a serious health issue in Indonesia, particularly among families with low socio-economic conditions. However, the lack of precise criteria or measurements of social conditions contributing to at-risk families makes prediction challenging. This study aims to identify patterns of relationships among 17 criteria influencing stunting risk, such as maternal age, number of children, type of flooring in the house, and access to clean water, by enhancing the efficiency of the Apriori algorithm through hash-based techniques. Data were obtained from families in Tuah Madani District, Pekanbaru, and analyzed using data preprocessing and transformation methods. The implementation of this algorithm within a web-based information system enables rapid and efficient analysis to identify stunting risks based on relevant combinations of criteria. The analysis results indicate that certain criteria, such as maternal age above 35 years, status as a couple of childbearing age (PUS), and having more than three children, are significantly associated with stunting risk, with a support value of 37.54% and a confidence level of 83.16%. This study contributes to the development of efficient methods for stunting risk analysis and provides a foundation for more targeted health interventions. Future researchers are advised to expand the data scope by including additional regions and different time periods to improve result generalization. Furthermore, incorporating other variables, such as maternal nutritional status or the education level of household heads, may offer deeper insights into understanding stunting risk patterns.
Penerapan Metode ADASYN Dalam Mengatasi Imbalanced Data Untuk Klasifikasi Penyakit Stroke Menggunakan Support Vector Machine Alwaliyanto; Siska Kurnia Gusti; Iis Afrianty; Fadhilah Syafria
Bulletin of Computer Science Research Vol. 5 No. 4 (2025): June 2025
Publisher : Forum Kerjasama Pendidikan Tinggi (FKPT)

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.47065/bulletincsr.v5i4.612

Abstract

Stroke is one of the leading causes of death and disability worldwide, making it essential to develop classification models that can assist in early and accurate diagnosis. This study aims to implement the Support Vector Machine (SVM) algorithm with three types of kernels linear, polynomial, and Radial Basis Function (RBF) to classify stroke disease data. The Adaptive Synthetic Sampling (ADASYN) method is employed to address the class imbalance problem, while model training and evaluation are carried out using 5-Fold Cross-Validation to ensure stable and reliable results. The findings indicate that ADASYN successfully improves the model’s sensitivity to stroke cases (the minority class), as reflected by an increase in recall and F1-score, despite a slight decrease in overall accuracy a common trade-off in handling imbalanced data. The linear kernel (after ADASYN) achieved the best performance after imbalance handling, with an average AUC-ROC of 0.8333, recall of 0.7827, and F1-score of 0.2181 for the stroke class. Although the F1-score remains relatively low, it improved compared to the pre-ADASYN results, indicating better detection of stroke cases. The implementation was conducted using Google Colab, which also contributed to efficient data processing and visualization. Overall, the results demonstrate that the combination of SVM and ADASYN is effective in enhancing the model’s sensitivity to minority classes and is well-suited for medical data classification tasks, particularly in the early diagnosis of stroke using machine learning approaches.
EVALUASI PERBANDINGAN PERFORMANSI LVQ 1, LVQ 2, DAN LVQ 3 DALAM KLASIFIKASI JENIS KELAMIN MENGGUNAKAN TULANG TENGKORAK DARMILA; IIS AFRIANTY; SUWANTO SANJAYA; RAHMAD ABDILLAH; IWAN ISKANDAR; FADHILAH SYAFRIA
Jurnal INSTEK (Informatika Sains dan Teknologi) Vol 7 No 2 (2022): OCTOBER
Publisher : Department of Informatics Engineering, Faculty of Science and Technology, Universitas Islam Negeri Alauddin, Makassar, Indonesia

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.24252/instek.v7i2.32659

Abstract

Klasifikasi merupakan teknik pengelompokkan data sesuai dengan karakteristik data yang telah ditentukan. Hasil performansi akurasi dapat menjadi ukuran keakuratan metode yang digunakan dalam proses klasifikasi. Teknik pengambilan data yang tidak sesuai dapat mengurangi hasil akurasi. Pada penelitian ini menggunakan metode Learning Vector Quantization (LVQ) 1, 2, dan 3 untuk melihat keakuratan metode klasifikasi dengan menggunakan teknik pengambilan data sampling. Data yang digunakan merupakan data pengukuran tulang tengkorak laki-laki dan perempuan yang berjumlah 2524 data. Pada LVQ 1 mendapatkan akurasi terbaik yaitu 91.39% dengan learning rate 0.1, 0.4, 0.7, 0.9. LVQ 2 mendapatkan akurasi terbaik 77.05% dengan learning rate 0.9 dan window 0.2. LVQ 3 mendapatkan akurasi terbaik yaitu 80.04% dengan learning rate 0.7, window 0.1, dan epsilon 0.3. Hal ini menunjukkan bahwa LVQ 1 lebih tepat untuk diterapkan terhadap multi-fitur pada dataset William W. Howells Craniometric dibandingkan LVQ 2 dan LVQ 3.
PERBANDINGAN PERFORMANSI DENGAN METODE CORRELATION BASED FEATURE SELECTION PADA LVQ 2 SURYA ADITYA GD; IIS AFRIANTY; SUWANTO SANJAYA; RAHMAD ABDILLAH; LESTARI HANDAYANI; FITRI INSANI
Jurnal INSTEK (Informatika Sains dan Teknologi) Vol 8 No 1 (2023): APRIL
Publisher : Department of Informatics Engineering, Faculty of Science and Technology, Universitas Islam Negeri Alauddin, Makassar, Indonesia

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.24252/instek.v8i1.37301

Abstract

Melakukan sebuah penelitian diperlukannya mengidentifikasi sebuah data yang sesuai dengan melakukan sebuah klasifikasi. Pengaruh dalam mendapatkan hasil akurasi yang maksimal dengan menentukan teknik penelitian secara tepat melalui proses klasifikasi. Pada penelitian ini melakukan perbandingan peningkatan performansi akurasi akurasi LVQ 2 dengan mengimplementasikan Correlation Based Feature Selection (CFS) pada dataset bertujuan keakuratan pengambilan data sampel dengan metode klasifikasi. Data parameter tulang tengkorak yang digunakan yaitu data pria dan wanita dengan jumlah data 2524 dan fitur 82. Penelitian LVQ 2 tanpa CFS dengan nilai learning rate (α) = 0.9 dan window 0.2 yang akurasi tertingginya memperoleh sebesar 77.05%, dan menggunakan CFS pada nilai α = 0.9 dan window = 0.3 hasil akurasi tertinggi yaitu 82,51%. Hal ini menunjukkan bahwa LVQ 2 menggunakan CFS sangat direkomendasikan baik dari segi performansi terhadap pada dataset Tengkorak dibandingkan LVQ 2 tanpa menggunakan CFS.
KOMPARASI METODE K-NEAREST NEIGHBORS DAN LONG SHORT TERM MEMORY PADA KLASIFIKASI TERJEMAHAN AL-QUR’AN Nurul Fatiara; Nazruddin Safaat H; Surya Agustian; Yusra; Iis Afrianty
ZONAsi: Jurnal Sistem Informasi Vol. 6 No. 2 (2024): Publikasi Artikel ZONAsi: Periode Mei 2024
Publisher : Universitas Lancang Kuning

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.31849/zn.v6i2.19863

Abstract

Al-Qur’an merupakan kitab suci yang diturunkan untuk umat islam. Secara harfiah, Al-Qur'an berasal dari kata qara’a yang artinya membaca atau mengumpulkan. Namun untuk memahami terjemahan dari Al-Qur’an tidaklah mudah. Salah satu cara yang dapat dilakukan untuk memahami dan mempelajarinya adalah melakukan klasifikasi terhadap terjemahan ayat Al-Qur’an. Penelitian ini mengklasifikasikan terjemahan Al-Qur'an bahasa Indonesia ke enam kelas yang berbeda. Metode yang digunakan adalah K-Nearest Neighbor (KNN) dan Long Short Term Memory (LSTM) dan membandingkan kedua metode untuk mendapatkan hasil performa klasifikasi yang tertinggi. Hasil klasifikasi menunjukkan model LSTM menghasilkan performa klasifikasi yang lebih tinggi yaitu berupa rata-rata F1-Score sebesar 65% dan rata-rata accuracy 96% dibandingkan model KNN dengan nilai rata-rata F1-Score sebesar 55% dan rata-rata accuracy 93%.
Perbandingan Teknik Penyeimbang Kelas Pada Multi-Layer Perceptron (MLP) Berbasis Backpropagation Untuk Klasifikasi Diabetes Mellitus Robby Azhar; Siska Kurnia Gusti; Iis Afrianty; Elvia Budianita
Bulletin of Computer Science Research Vol. 5 No. 6 (2025): October 2025
Publisher : Forum Kerjasama Pendidikan Tinggi (FKPT)

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.47065/bulletincsr.v5i6.804

Abstract

Diabetes Mellitus (DM) is a chronic disease that can lead to serious complications if not detected early; therefore, early diagnosis is highly important. One of the methods that can be applied for early diagnosis is the classification technique in data mining. However, the classification process often faces challenges due to class imbalance, which can reduce model performance. This study aims to analyze the effect of class balancing techniques on the performance of the Backpropagation Neural Network (BPNN) in classifying DM cases. BPNN is a form of Multi-Layer Perceptron (MLP) with a simple structure and the ability to solve complex problems with good accuracy. The dataset used in this study is the Pima Indians Diabetes Dataset, consisting of 768 instances, including 500 non-diabetic and 268 diabetic cases. The research was conducted using three scenarios: without balancing, Synthetic Minority Over-sampling Technique (SMOTE), and Random Under Sampling (RUS). The BPNN model was designed with two architectural variations (one hidden layer and two hidden layers), three learning rate values (0.1, 0.01, and 0.001), and a varying number of neurons. The dataset was divided using the 10-Fold Cross Validation technique. The results show that applying SMOTE achieved the best performance, with an average accuracy of 90.89%, precision of 91.22%, recall of 90.89%, and F1-score of 90.89% on the BPNN architecture with one hidden layer. Furthermore, the single hidden layer architecture proved more stable than the two hidden layers, especially when the dataset size decreased due to RUS. Therefore, the combination of SMOTE and BPNN with one hidden layer provides better performance in classifying Diabetes Mellitus cases.