Comparison of artificial neural network developed in matlab and python programming language

dc.contributor.authorMuhammad Zulfaqar Bin Mohd Roslan
dc.date.accessioned2021-03-04T05:13:06Z
dc.date.available2021-03-04T05:13:06Z
dc.date.issued2019-06
dc.description.abstractNeural network is neurons that are connected by synapses. Neural network consist of three main layers which are input layer, hidden layer and output layer. It undergoes deep learning process to produce higher accuracy of predicted output. The neural network used sigmoid function as activation function because it is able to map the results ranging from 0 to 1. Mean sum squared loss function was used to calculate the error by altering the weights on each iteration until the error is approaching zero. The neural network is design by using MATLAB and Python programming language. Third party software is imported to help in calculation of artificial neural network in Python programming language such as GEKKO and NumPy. Neural network in Python programming language consisting of 2 nodes of input layers, 10 nodes of hidden layers and 1 node of output layer that was run for 500 iterations and the same model are build in MATLAB. Neural network in Python programming language showed 1.030 × 10 -3 error while 1.401 × 10 -9 error are shown in MATLAB which both can be considered nearly equivalent to 0. The neural network created was then tested on shell and tube heat exchanger. The flow rate of cold water stream inlet and the temperature of hot water stream inlet were used as the input for the neural network. The neural network was able to predict the output which was deviated 1.95℃ the most from the actual output during training process while it deviate 2.03℃ the most during test process.en_US
dc.identifier.urihttp://hdl.handle.net/123456789/11766
dc.language.isoenen_US
dc.titleComparison of artificial neural network developed in matlab and python programming languageen_US
dc.typeOtheren_US
Files
License bundle
Now showing 1 - 1 of 1
Loading...
Thumbnail Image
Name:
license.txt
Size:
1.71 KB
Format:
Item-specific license agreed upon to submission
Description: