Claim Missing Document
Check
Articles

Found 2 Documents
Search

IMPLEMENTASI NEURAL NETWORK BACKPROPAGATION UNTUK IDENTIFIKASI TINGKAT MANIS BUAH BELIMBING BERDASARKAN CITRA RGB Retno Nugroho Whidhiasih
PIKSEL : Penelitian Ilmu Komputer Sistem Embedded and Logic Vol. 2 No. 1 (2014): Maret 2014
Publisher : LPPM Universitas Islam 45 Bekasi

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

Abstract

ABSTRACT Star fruit classification is needed to maintain quality and improve competitiveness. Star fruit-based sweetness can be done destructively and non-destructively. Nondestructive can be done by measuring the correlation value of red, green, blue (RGB) star fruit image with Total Dissolved Solids (TPT) contained in starfruit. This study aims to develop an artificial intelligence system model to classify star fruit non-destructively based on the red-green-blue component using Neural Network (NN). The input parameter used is the red-green-blue component of the star fruit image which has been correlated to the TPT. The amount of sample data used is 99 pieces, which is 33 sweet starfruit image, 33 medium starfruit image and 33 image starfruit acid. A total of 81 data were used as training data and 18 data were used as test data. To obtain the best introductory results experiments were conducted using 6 variations of the number of neurons in the hidden layer. The classification into acid, medium and sweet fruit classes in this study obtained the best NN model using red, green and blue input parameters with 2 neurons in the hidden layer. The NN backpropatation 3-2-1 model provides an accuracy of 66.67% with 2 neurons in the hidden layer, MSE of 4.73e-06 on epoch 1. Keyword : classification, neural network, starfruit, non-destructive grading, pattern recognition. ABSTRAK Pemutuan buah belimbing sangat diperlukan untuk mempertahankan mutu dan meningkatkan daya saing. Pemutuan buah belimbing berdasarkan rasa dapat dilakukan secara destruktif dan nondestruktif. Nondestruktif dapat dilakukan dengan mengukur korelasi nilai red, green, blue (RGB) citra buah belimbing dengan Total Padatan Terlarut (TPT) yang terdapat pada belimbing. Penelitian ini bertujuan untuk mengembangkan model sistem kecerdasan buatan untuk mengklasifikasi buah belimbing secara non-destruktif berdasarkan komponen red-green-blue menggunakan Neural Network (NN). Parameter input yang digunakan adalah komponen red-green-blue dari citra buah belimbing yang telah dikorelasikan terhadapTPT. Jumlah sampel data yang digunakan adalah 99 buah, yaitu 33 citra belimbing manis, 33 citra belimbing sedang dan 33 citra belimbing asam. Sejumlah 81 data digunakan sebagai data pelatihan dan 18 data digunakan sebagai data pengujian. Untuk mendapatkan hasil pengenalan terbaik dilakukan percobaan-percobaan menggunakan 6 variasi jumlah neuron pada lapisan tersembunyi. Klasifikasi menjadi kelas buah asam, sedang dan manis dalam penelitian ini mendapatkan model NN terbaik menggunakan parameter input red, green dan blue dengan 2 neuron pada lapisan tersembunyi. Model NN backpropatation 3-2-1 ini memberikan akurasi sebesar 66.67% dengan 2 neuron pada lapisan tersembunyi, MSE sebesar 4.73e-06 pada epoch ke 1. Keyword : klasifikasi, belimbing, neural network, citra digital, pemutuan non-destruktif
Evaluasi Penggunaan Website Kampus Menggunakan Metode Usability Testing dan System Usability Scale Inna Ekawati; Malikus Sumadyo; Retno Nugroho Whidhiasih
JREC (Journal of Electrical and Electronics) Vol. 11 No. 2 (2023): JREC (Journal of Electrical and Electronics)
Publisher : Program Studi Teknik Elektro Fakultas Teknik Universitas Islam 45 Bekasi

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.33558/jrec.v11i2.7957

Abstract

The campus website is the face of an institution that can be accessed online by everyone because the website is a central information service. When individuals such as potential students, the wider public, and those within the institution itself require information about the campus, they will primarily turn to the website as their main source. However, not all campus websites have a perfect role as information and service centers. The quality of a campus website can be seen from the perception and feedback of its users. Therefore, an assessment of online media usage, especially websites, is essential at all times to understand user perceptions of the ongoing website development. This research is intended to evaluate the usability of the website of Islamic University 45 Bekasi to obtain the best recommendations for improvement. In addition to the evaluation, this research also compares two methods of website usability evaluation. The methods to be used are usability testing and the System Usability Scale. Both methods are conducted simultaneously to obtain qualitative and quantitative data. This research yields recommendations for areas that need improvement or enhancement and also provides a comparison of the two methods used as references for usability evaluation. The research results reveal that with the first method, there is a significantly higher number of positive perceptions than negative ones, although there are many suggestions for improvement. On the other hand, when using the second method, the campus website exhibits the second least favorable usability score among a total of eleven levels.