Claim Missing Document
Check
Articles

Found 18 Documents
Search

Face pattern recognition using Expectation-Maximization (EM) algorithm Purwadi, Joko; Hernadi, Julan; Suryantoro, M. Danang
Bulletin of Applied Mathematics and Mathematics Education Vol. 2 No. 1 (2022)
Publisher : Universitas Ahmad Dahlan

Show Abstract | Download Original | Original Source | Check in Google Scholar | Full PDF (824.735 KB) | DOI: 10.12928/bamme.v2i1.5520

Abstract

This paper discuss about the use face patteren recognition which is now days become popular especialy on smartphone lock screen system. The method used in this research are the Expectation – Maximization (EM) Algorithm. EM Algorithm is an iterative optimization method for the estimation of Maximum Likelihood (ML) which is used in incomplete data problems. there are 2 stages, namely the Expectation stage E (E-step) and the Maximization stage M (M-step). These two stages will continue to be carried out until they reach a convergent value. The result of the research shows that EM Algorthm produce high accuracy, it’s about 95% on the data training and 83% accuracy on the data testing.
Support Vector Regression optimization with Particle Swam Optimization algorithm for predicting the gold prices Selviani, Novi; Purwadi, Joko
Bulletin of Applied Mathematics and Mathematics Education Vol. 3 No. 2 (2023)
Publisher : Universitas Ahmad Dahlan

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.12928/bamme.v3i2.9561

Abstract

This paper discusses about how to predict the gold prices from 1 January 2021 to 31 January 2023. The method used in this study is the Support Vector Regression (SVR) technique, method that was developed from the support vector machine which is used as regression approach to predict future event. From the past study already know that SVR had limitation in achieving good performance because of its sensitivity to parameters. To overcome the SVR performance problems, an optimization algorithm is proposed in this study. The PSO algorithm is applied in this study to optimize the parameters of the SVR method. The results showed that the prediction of the SVR model obtained an MSE value of 0.0035744. While in the SVR model with the PSO algorithm, the MSE value is 0.0033058.
Optimasi Parameter Support Vector Regression (SVR) Menggunakan Algoritma Grey Wolf Optimizer (GWO) Yulia Candra Dewi; Joko Purwadi
Jurnal Ilmiah Matematika Vol. 10 No. 1 (2023)
Publisher : Universitas Ahmad Dahlan

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.26555/jim.v10i1.30867

Abstract

Prediksi harga bawang merah merupakan hal penting bagi petani dan pemerintah untuk mengurangi risiko ekonomi dan membuat keputusan yang lebih baik. Penelitian ini bertujuan untuk mengembangkan model prediksi harga bawang merah di Indonesia menggunakan Support Vector Regression (SVR) yang dioptimalkan dengan algoritma Grey Wolf Optimizer (GWO). SVR adalah teknik pembelajaran mesin yang efektif untuk regresi, tetapi mempunyai kesulitan dalam menetapkan parameter optimalnya. Untuk itu, algoritma GWO, yang terinspirasi dari strategi berburu serigala, digunakan untuk mengoptimalkan parameter SVR. Dalam penelitian ini, data harga bawang merah sejak tanggal 1 Januari 2022 sampai 31 Desember 2023 yang diperoleh dari website resmi Pusat Informasi Harga Pangan Strategis Nasional (PIHPS) dikumpulkan dan dianalisis. Hasil penelitian menunjukkan bahwa tingkat eror yang diukur dengan RMSE (Root Mean Square Error) untuk model GWO-SVR diperoleh sebesar 0.062561 sedangkan model SVR sebesar 0.078579. Dapat dilihat bahwa terjadi penurunan nilai RMSE, sehingga dapat dikatakan bahwa algoritma optimasi GWO dapat meningkatkan kinerja dari model SVR.
Implementasi Metode SVM-PSO Dengan Fitur Selection Variance Threshold Pada Klasifikasi Penyakit Diabetes Mellitus Pratiwi Kistiya Ningrum; Joko Purwadi
Jurnal Ilmiah Matematika Vol. 10 No. 2 (2023)
Publisher : Universitas Ahmad Dahlan

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.26555/jim.v10i2.30877

Abstract

Pada penelitian ini membahas tentang kasus klasifikasi pada data penyakit diabetes. Metode yang digunakan dalam penelitian ini adalah metode Support Vector Machine yang dioptimalkan dengan algoritma Particle Swarm Optimization guna memperoleh parameter terbaik dengan kombinasi seleksi fitur menggunakan Variance Threshold. Penelitian ini bertujuan untuk mengetahui cara kerja dan hasil akurasi dari metode Support Vector Machine dengan optimasi Particle Swarm Optimization menggunakan seleksi fitur Variance Threshold. Hasil penelitian menggunakan kombinasi metode tersebut menunjukkan hasil akurasi sebesar 80%. Hasil akurasi tersebut lebih tinggi jika dibandingkan dengan metode Support Vector Machine tunggal tanpa optimasi dan seleksi fitur dengan akurasi sebesar 76%. Meningkatkan akurasi sebesar 4% dari 76% menjadi 80%.
ANALISIS PENGENDALIAN KUALITAS PRODUKSI BERAS DENGAN METODE STATISTICAL PROCESS CONTROL (SPC) Luthfi Alleyda Fadhlullah; Joko Purwadi
Jurnal Ilmiah Matematika Vol. 11 No. 2 (2024)
Publisher : Universitas Ahmad Dahlan

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.26555/jim.v11i2.30893

Abstract

Perkembangan industri meningkat dikarenakan kebutuhan manusia yang beranekaragam seperti bahan pangan, papan, sandang, dan kendaraan. Perkembangan ini mendorong perusahaan yang bergeras dibidang industrialisasi untuk terus menjaga bahkan meningkatkan kualitas produk yang mereka hasilkan untuk menjaga kepercayaan pelanggan. UD. Penggilingan X merupakan bidang usaha yang bergerak dibidang industri pangan yang memproduksi Beras. Beras merupakan salah satu produk makanan pokok paling penting di dunia, termasuk di Indonesia. Pada penggilingan diperlukan penjagaan kualitas agar nantinya beras yang dihasilkan akan selalu terjaga bahkan meningkat setiap harinya. Kualitas ini dapat dijaga dengan ilmu matematis yaitu pengendalian kualitas yakni menggunakan tujuh alat Statistical Process Control (SPC).
Support Vector Machine for Classification: A Mathematical and Scientific Approach in Data Analysis Yulia Restiani; Joko Purwadi
Jurnal Penelitian Pendidikan IPA Vol 10 No 11 (2024): November
Publisher : Postgraduate, University of Mataram

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.29303/jppipa.v10i11.8122

Abstract

In this research, SVM will be used to differentiate between plain nail art designs (class 0), 3D nail art designs (class 1), and hand painting nail art designs (class 2). The dataset used consists of images of nail designs that have been collected and analyzed previously. First, the dataset is divided into three different classes based on the type of nail design. The first class (class 0) includes plain nail art designs, then the second class (class 1) is 3D nail art designs, and the third class (class 2) is hand painting nail art designs. This process is carried out to allow SVM to learn the feature differences between the two types of designs. The data used will be divided into training and testing data and divided into three data division schemes, namely 60/40, 70/30, and 80/20. Based on the results of the research discussed, it can be concluded that classification using the Linear SVM model on three data sharing schemes provides the best level of accuracy on the 80/20 scheme, namely 81.25%. Meanwhile, classification using the non-linear SVM model achieved the highest level of accuracy of 95% in the 80/20 scheme with the RBF Kernel. Thus, the SVM model that is suitable for classifying nail art designs is a non-linear SVM model with the 80/20 scheme. The accuracy results obtained from this research also show that SVM provides good performance in classifying nail art designs.
Spatial Regression Analysis using Queen Contiguity Weight Matrix and PCA Dimensionality Reduction Joko Purwadi; Iliana Dewinta
CAUCHY: Jurnal Matematika Murni dan Aplikasi Vol 11, No 1 (2026): CAUCHY: JURNAL MATEMATIKA MURNI DAN APLIKASI
Publisher : Mathematics Department, Maulana Malik Ibrahim State Islamic University of Malang

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.18860/cauchy.v11i1.40908

Abstract

Conventional linear regression often falls short in poverty analysis, as it fails to account for spatial interdependence between neighboring regions and frequently encounters multicollinearity among socioeconomic variables. This study investigates the presence and nature of spatial effects in poverty data across regencies and cities in Central Java Province, Indonesia, and assesses the performance of an enhanced spatial regression model. We employ a Spatial Autoregressive Model (SAR) integrated with a queen contiguity spatial weight matrix and apply Principal Component Analysis (PCA) to reduce dimensionality and mitigate multicollinearity. The results demonstrate a strong model fit, with a pseudo R2 of 0.94311, and reveal a statistically significant negative spatial lag coefficient ( = -0.2039, p-value = 0.04420), indicating that areas of lower poverty are often surrounded by higher poverty neighbors. This integrated approach provides a more accurate framework for spatial poverty mapping, offering actionable insights for designing regionally targeted development policies.
Spectral Clustering-Based Segmentation Framework for TikTok Influencer Classification Rizky Ageng Saputra; Joko Purwadi
ZERO: Jurnal Sains, Matematika dan Terapan Vol 9, No 2 (2025): Zero: Jurnal Sains Matematika dan Terapan
Publisher : UIN Sumatera Utara

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.30829/zero.v9i2.26396

Abstract

This study presents a data-driven segmentation model for TikTok influencers using Spectral Clustering on 120 verified beauty influencers from FastMoss TikTok Analytics (2024-2025). Five engagement metrics views, likes, comments, shares, and followers were selected via variance thresholding, explaining 92.6% of behavioral variance. A similarity graph with a Radial Basis Function (RBF) kernel (σ = 0.5) and k = 3 clusters yielded a Silhouette Score of 0.9473, indicating highly cohesive and well-separated clusters. Compared to K-Means and Hierarchical Clustering, Spectral Clustering achieved 7.8% higher cohesion, capturing complex, nonlinear engagement patterns. Principal Component Analysis (PCA) confirmed clear distinctions among Micro-Mid, Macro, and Mega influencers. Results show that influencer impact depends more on interaction dynamics than follower count, offering a graph-based approach to optimize brand strategies effectively.