Claim Missing Document
Check
Articles

Found 2 Documents
Search

Analisis Sentimen terhadap Kebijakan Pengembalian Penjurusan di Sekolah Menengah Atas Menggunakan Algoritma Random Forest Saire Fernando; Mulyono; Ari Hendarno; Ersa Resita; Carli Apriansyah Hutagalung
J-KOMA : Jurnal Ilmu Komputer dan Aplikasi Vol 9 No 01 (2026): J-KOMA : Jurnal Ilmu Komputer dan Aplikasi
Publisher : Universitas Negeri Jakarta

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.21009/JKOMA.091.02

Abstract

In the 2025/2026 academic year, the Minister of Education, Culture, Research, and Technology (Permendikbudristek) planned to reinstate majors in high schools. This policy quickly sparked public debate, generating a variety of responses from the public, particularly on YouTube. This study was conducted to analyze public sentiment towards this policy using the Random Forest algorithm. Data was obtained through YouTube comment collection between April 2025 and June 2025, which yielded 4,181 comments. The analysis showed that YouTube comments on videos regarding the policy of reinstating majors in the high school curriculum had a more positive response than negative, with a ratio of 18.3% compared to 5%, with a total of 76.7% neutral. The model's performance in classifying sentiment achieved an overall accuracy of 83.16%, with a precision of 77.56%, indicating a fairly reliable model prediction. However, recall was at 57.25%, indicating that the model still had difficulty detecting all data in small classes such as positive and negative. Overall, the model's performance is represented by an F1-score of 61.26%. This research is expected to serve as a consideration for education leaders or the Minister of Education, Culture, Research, and Technology in determining whether a policy needs to be enacted or changed.
PERFORMANCE ANALYSIS OF OPTICAL CHARACTER RECOGNITION (OCR) METHOD FOR VEHICLE LICENSE PLATE IDENTIFICATION Ersa Resita; Dinar Munggaran Akhmadi; Carli Apriansyah Hutagalung; Renaldo Yusellino; Zhafirah Naswa N
JOURNAL OF SCIENCE AND SOCIAL RESEARCH Vol. 9 No. 5 (2026): October (2026)
Publisher : Smart Education

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.54314/dcahzk48

Abstract

Abstract: The increasing number of motor vehicles has created a growing need for accurate and efficient vehicle license plate recognition systems. Conventional manual identification methods are often time-consuming and prone to human error. This study aims to evaluate the performance of Optical Character Recognition (OCR) in recognizing vehicle license plate characters through digital image processing techniques. The proposed approach involves several preprocessing stages, including image resizing, grayscale conversion, denoising, and thresholding, to improve image quality before character recognition. The license plate region is identified using a bounding box, and the extracted region is processed by the OCR system to recognize alphanumeric characters. System performance is evaluated using character recognition accuracy, character error rate, and processing time across a collection of vehicle license plate images with varying image quality and lighting conditions. The evaluation results are expected to provide insights into the effectiveness and limitations of OCR in vehicle license plate recognition and identify factors that influence recognition performance. This study is expected to contribute to the development of reliable automated license plate recognition systems for applications such as smart parking and vehicle access management. Keyword: Smart Parking, OCR, Deep Learning, Number Plates, Computer Vision.   Abstrak: Meningkatnya jumlah kendaraan bermotor telah menciptakan kebutuhan yang semakin besar akan sistem pengenalan plat nomor kendaraan yang akurat dan efisien. Metode identifikasi manual konvensional seringkali memakan waktu dan rentan terhadap kesalahan manusia. Studi ini bertujuan untuk mengevaluasi kinerja Pengenalan Karakter Optik (OCR) dalam mengenali karakter plat nomor kendaraan melalui teknik pengolahan citra digital. Pendekatan yang diusulkan melibatkan beberapa tahap pra-pemrosesan, termasuk pengubahan ukuran citra, konversi skala abu-abu, pengurangan noise, dan penentuan ambang batas, untuk meningkatkan kualitas citra sebelum pengenalan karakter. Wilayah plat nomor diidentifikasi menggunakan kotak pembatas, dan wilayah yang diekstrak diproses oleh sistem OCR untuk mengenali karakter alfanumerik. Kinerja sistem dievaluasi menggunakan akurasi pengenalan karakter, tingkat kesalahan karakter, dan waktu pemrosesan pada kumpulan citra plat nomor kendaraan dengan kualitas citra dan kondisi pencahayaan yang bervariasi. Hasil evaluasi diharapkan dapat memberikan wawasan tentang efektivitas dan keterbatasan OCR dalam pengenalan plat nomor kendaraan dan mengidentifikasi faktor-faktor yang memengaruhi kinerja pengenalan. Studi ini diharapkan dapat berkontribusi pada pengembangan sistem pengenalan plat nomor otomatis yang andal untuk aplikasi seperti parkir pintar dan manajemen akses kendaraan. Kata kunci: Smart Parking, OCR, Deep Learning, Plat Nomor, Computer Vision