Article Per Year (5 Year)
p-Index From 2020 - 2025
0.23
P-Index
This Author published in this journals
All Journal INFOKUM
Maranata Pasaribu
Universitas Mandiri Bina Prestasi (MBP)
Computer Science & IT

Published : 1 Documents Claim Missing Document
Claim Missing Document
Check
Articles

Found 1 Documents
Search

 1 
Analysis of Elearning Quality Measurement with Webqual method using Artificial Neural Networks Erwin Daniel Sitanggang; Misdem Sembiring; Anjar Pinem; Maranata Pasaribu
INFOKUM Vol. 10 No. 02 (2022): Juni, Data Mining, Image Processing, and artificial intelligence
Publisher : Sean Institute

Show Abstract | Download Original | Original Source | Check in Google Scholar | Full PDF (469.138 KB)

Abstract

Currently, artificial intelligence is a concern for the world because of its increasingly rapid and sophisticated application in helping humans to complete their work in everyday life. One of the widely used methods is artificial neural networks that are part of deep learning and a subsection of machine learning. In its network training, the data used as input is the gap score of each webqual dimension and the data used as the output is the gap score of the average webqual attributes of each respondent. The training process is expected to produce an actual output close to the predetermined target output, resulting in the best model of artificial neural networks with feedforward backpropagation algorithms. From the results of the training experiment, the best model of artificial neural network architecture was obtained with a feedforward backpropagation algorithm at the time of training from 174 data to be able to replace the Webqual method in this study using the 3-20-1 model and the algorithm used was Levenberg-Marquardt (trainln). Where there is 1 Input layer with 3 neuron units, 1 hidden layer with 20 neuron units and 1 Output layer with 1 neuron unit with a mean square error (mse) of 0.00000000000721 and regression of 1 or 100%. And after testing using 58 data using the network configuration obtained during training, the results of the comparison between the network output and the target were 100% accurate.
Co-Authors Anjar Pinem Erwin Daniel Sitanggang Misdem Sembiring