Journal of Computing and Informatics Research

Signature Identification Based on GLCM Feature Extraction and Convolutional Neural Network Classification

2026-07-15T04:05:16+00:00

Signature is one of the widely used biometric features for authentication and identity verification purposes. This study proposes a digital signature identification system by combining Gray Level Co-occurrence Matrix (GLCM) feature extraction and Convolutional Neural Network (CNN) classification. The dataset consists of 500 signature images from 50 individuals collected independently. Preprocessing steps include grayscale conversion, adaptive binarization, and image normalization to 128×128 pixels. GLCM texture features are extracted at four angular directions (0°, 45°, 90°, 135°) yielding five main features: contrast, correlation, energy, homogeneity, and entropy. These features are integrated as additional inputs to the fully connected layer of a CNN comprising three convolutional blocks. Experimental results demonstrate that the proposed system achieves a classification accuracy of 96.8%, precision of 96.2%, and F1-Score of 96.5% on test data. These results confirm that integrating GLCM texture features into the CNN architecture significantly improves signature identification performance.

Analisis Kondisi Sosial Ekonomi Santri Menggunakan Metode K-Means Clustering

2026-07-14T17:07:21+00:00

Abstrak− Analisis terhadap kondisi sosial santri di MI Al Huda Kota Malang disusun berdasarkan sejumlah indikator penting yang meliputi tingkatpenghasilan, jenis pekerjaan, latar belakang pendidikan, serta kepemilikantempat tinggal. Ketiadaan sistem klasifikasi yang terorganisir menyebabkanproses penentuan tingkat kesejahteraan belum dapat dilakukan secara objektifdan terukur. Untuk mengatasi hal tersebut, digunakan pendekatan data miningmelalui penerapan algoritma K-Means Clustering sebagai metode pengelompokan data. Tahapan awal dilakukan dengan proses preprocessing data, yang mencakup pemilihan variabel yang relevan, penanganan data, sertanormalisasi menggunakan StandardScaler agar distribusi data menjadi lebihseragam. Setelah itu, evaluasi terhadap hasil clustering dilakukan denganmemanfaatkan beberapa metrik, yaitu Elbow Method, Silhouette Score, dan Davies-Bouldin Index guna menentukan jumlah cluster yang optimal.Berdasarkan hasil evaluasi, diperoleh jumlah cluster terbaik sebanyak empat (k = 4), dengan nilai Silhouette Score sebesar 0,3341 dan Davies-Bouldin Indexsebesar 1,1277. Setiap cluster kemudian dianalisis lebih lanjut sesuaikarakteristik masing-masing tanpa melakukan perubahan terhadap jumlahcluster yang telah ditetapkan. Sebagai perbandingan, algoritma DBSCAN juga diterapkan, namun menunjukkan performa yang lebih rendah dibandingkandengan K-Means dalam menghasilkan kualitas pengelompokan. Secarakeseluruhan, metode K-Means terbukti mampu memberikan hasil clusteringyang lebih konsisten dan representatif, sehingga dapat dimanfaatkan sebagai dasar dalam pengambilan keputusan terkait kondisi sosial ekonomi santri.

Segmentasi Pelanggan E-Commerce Berbasis Perilaku Belanja Menggunakan Algoritma K-Means Clustering dan Evaluasi Silhouette Score

2026-07-29T11:02:48+00:00

Dalam era persaingan E-Commerce yang ketat, pemahaman mendalam mengenai perilaku konsumen menjadi kunci utama keberhasilan strategi pemasaran. Penelitian ini bertujuan untuk melakukan segmentasi pelanggan toko online menggunakan algoritma K-Means Clustering berbasis perilaku belanja. Menggunakan dataset Mall Customer Segmentation, penelitian ini memproses fitur utama berupa Pendapatan Tahunan dan Skor Pengeluaran yang dinormalisasi. Evaluasi model dilakukan menggunakan metode Elbow dan Silhouette Score untuk menentukan jumlah kelompok optimal. Hasil penelitian menunjukkan bahwa pembagian pelanggan menjadi 5 cluster adalah yang paling optimal dengan nilai Silhouette Score sebesar 0.5547. Kelima segmen yang terbentuk meliputi Middle Class (40,5%), VIP/Big Spenders (19,5%), Hemat Mapan (17,5%), Ekonomis (11,5%), dan Impulsif (11%). Temuan ini memberikan wawasan strategis bagi pelaku bisnis untuk merancang kampanye pemasaran yang terpersonalisasi, seperti program loyalitas eksklusif untuk segmen VIP dan penawaran diskon intensif untuk segmen Ekonomis, guna meningkatkan retensi dan profitabilitas

Optimasi Seleksi Fitur Adaptive Particle Swarm Optimization Untuk Klasifikasi Penyakit Jantung Dengan Ensemble Learning

2026-07-22T17:51:13+00:00

Heart disease classification using machine learning requires relevant features and predictive models capable of consistently generalizing clinical patterns. Previous studies on the Heart Failure Prediction dataset demonstrated that K-Nearest Neighbor (KNN) optimized with Particle Swarm Optimization (PSO) achieved an accuracy of 89.09% and an Area Under the Curve (AUC) of 0.935. However, the use of a fixed inertia weight and reliance on a single learner may limit the balance between exploration and exploitation, thereby reducing model robustness. This study proposes a feature selection approach based on Adaptive Particle Swarm Optimization (APSO), in which the inertia weight is gradually decreased from 0.90 to approximately 0.42 over 30 iterations. The optimal feature subset is subsequently utilized in a soft voting ensemble learning model. The dataset consists of 918 records, 11 predictive features, and one target class (HeartDisease). Experimental results indicate that the proposed APSO-based ensemble model achieved an accuracy of 89.71%, an F1-score of 0.8986, and an AUC of 0.9466. The confusion matrix yielded 90 true negatives, 12 false positives, 9 false negatives, and 93 true positives on 204 testing instances. Compared with the baseline KNN-PSO model, the proposed method improved classification accuracy by 0.62 percentage points and increased the AUC by 0.0116, while maintaining a disease-class recall of 91.18%. These findings demonstrate that combining adaptive search dynamics with heterogeneous ensemble learning enhances the discriminative capability of heart disease classification, although further validation using identical data partitioning strategies and external datasets is still required

Sistem Pendukung Keputusan Penerima BANPRES UMKM Penanganan Covid-19 Menerapkan Metode WASPAS

2026-07-29T11:12:40+00:00

The Covid-19 pandemic has had a significant impact on the sustainability of Micro, Small, and Medium Enterprises (MSMEs) in Indonesia. To assist affected MSMEs, the government launched the Presidential Assistance for Micro Business Productiveness (BANPRES UMKM) program. However, in its implementation, the process of determining aid recipients often faces obstacles such as inaccurate targeting, lack of objectivity, and time constraints in the selection process. Therefore, a system is needed that can assist in effective and efficient decision-making. This study aims to design a Decision Support System (DSS) to determine BANPRES UMKM recipients by applying the Weighted Aggregated Sum Product Assessment (WASPAS) method. The WASPAS method was chosen because it is able to integrate the advantages of the Weighted Sum Model (WSM) and Weighted Product Model (WPM) methods, resulting in more stable and accurate calculations in multi-criteria problems. The criteria used in this system include business legality, length of business operation, level of losses due to the pandemic, number of family dependents, and asset ownership. System testing results demonstrate that the WASPAS method is capable of objectively ranking alternatives, thus supporting a more transparent and targeted selection process. This research is expected to provide a technological solution to support data-driven social assistance distribution and measurable decision-making logic.