THE EFFECT OF THE NUMBER OF HIDDEN LAYERS IN THE BACKPROPAGATION IN CASE STUDY WEATHER CLASSIFICATION

VERAWATI, ANG, ESTER (2018) THE EFFECT OF THE NUMBER OF HIDDEN LAYERS IN THE BACKPROPAGATION IN CASE STUDY WEATHER CLASSIFICATION. Other thesis, Unika Soegijapranata Semarang.

[img]
Preview
Text (COVER)
14.K1.0017 Ang, Ester Verawati (2.39%).COVER.pdf

Download (294kB) | Preview
[img]
Preview
Text (BAB I)
14.K1.0017 Ang, Ester Verawati (2.39%).BAB I.pdf

Download (67kB) | Preview
[img] Text (BAB II)
14.K1.0017 Ang, Ester Verawati (2.39%).BAB II.pdf
Restricted to Registered users only

Download (64kB)
[img]
Preview
Text (BAB III)
14.K1.0017 Ang, Ester Verawati (2.39%).BAB III.pdf

Download (80kB) | Preview
[img]
Preview
Text (BAB IV)
14.K1.0017 Ang, Ester Verawati (2.39%).BAB IV.pdf

Download (195kB) | Preview
[img]
Preview
Text (BAB V)
14.K1.0017 Ang, Ester Verawati (2.39%).BAB V.pdf

Download (138kB) | Preview
[img]
Preview
Text (BAB VI)
14.K1.0017 Ang, Ester Verawati (2.39%).BAB VI.pdf

Download (58kB) | Preview
[img]
Preview
Text (DAFTAR PUSTAKA)
14.K1.0017 Ang, Ester Verawati (2.39%).DAFTAR PUSTAKA.pdf

Download (63kB) | Preview
[img]
Preview
Text (LAMPIRAN)
14.K1.0017 Ang, Ester Verawati (2.39%).LAMPIRAN.pdf

Download (74kB) | Preview

Abstract

ABSTRACT Weather is something that is important in human life. Using a variety of applications the development of modern times, the weather can be predicted, for human needs. But often becomes a problem is the prediction accuracy of the weather classification. Because with these predictions, many people make weather predictions to plan daily activities as well as in the near future, such as vacation plans, determining which transportation to travel to, etc. The ANN algorithm is an algorithm that is like the workings of a human neuron in sending an impulse. Backpropagation is one of the ANN algorithms that can be used for classification. Backpropagation is done through two stages of the process of learning and testing. It is requires learning process to learn the pattern of output to be generated, and testing process to classify the results of learning process. Backpropagation has three layers, there arethe input layer, the hidden layer and output layer. Input layer is a layer that contains the input parameters of the data to be processed. Hidden layer is a layer that contains a number of values of the input parameters that have been processed using a number of weights. The output layer is the layer that contains the output value generated. This project examines the effect of the number of hidden layer on the accuracy of weather classification. Based on the results of testing in this research, there is no significant change in the accuracy of weather classification results in backpropagation with 1, 2 and 3 hidden layers. However, backpropagation with more number of hidden layers need more epoch. Also the error value calculated using the MSE (Mean Square Error) calculation becomes smaller with the more number of hidden layers in the backpropagation. Keyword: backpropagation, weather classification, hidden layer

Item Type: Thesis (Other)
Subjects: 000 Computer Science, Information and General Works > 005 Computer programming, programs & data > Information Systems
Divisions: Faculty of Computer Science > Department of Informatics Engineering
Depositing User: Mr Lucius Oentoeng
Date Deposited: 21 Jun 2018 04:25
Last Modified: 21 Jun 2018 04:25
URI: http://repository.unika.ac.id/id/eprint/16179

Actions (login required)

View Item View Item