Optimal configuration of multilayer perceptron neural network classifier for recognition of intracranial epileptic seizures

A classification problem in high energy physics has been solved on simulated data using a simple multilayer perceptron comprising binary units which was trained with the CHIR algorithm. The unstable training of such a network on a nonseparable set has been overcome by selecting those weight vectors with good performance while providing a flexible choice of the two types of classification errors. Specific features of the problem have been exploited in order to simplify and optimize the solution which has been compared to the popular backpropagation algorithm and found to perform on a similar level. Additional aspects of this work are the use of the CHIR algorithm on continuous input and incorporating the classic idea of a Φ-machine in a multilayer perceptron.

Download Full-text

Classification of economic data into multiple classes by means of evolutionary methods

Acta Universitatis Agriculturae et Silviculturae Mendelianae Brunensis ◽

10.11118/actaun201361072445 ◽

2013 ◽

Vol 61 (7) ◽

pp. 2445-2449

Author(s):

Jiří Lýsek ◽

Jiří Šťastný

Keyword(s):

Neural Network ◽

Czech Republic ◽

Multilayer Perceptron ◽

Consumer Behaviour ◽

Neural Network Classifier ◽

Economic Data ◽

Kohonen Neural Network ◽

Grammar Rules ◽

Food Market

In this contribution we deal with an automatic classification of economic data into multiple classes. A classifier created by grammatical evolution is used to determine the data sample membership into one of the defined classes. The grammar rules used for classifier structure creation are presented. The performance of our classifier is compared with multilayer perceptron neural network classifier and Kohonen neural network classifier. We used a survey data of consumer behaviour in food market in Czech Republic.

Download Full-text

Prediction of Zika Virus by Multilayer Perceptron Neural Network (MLPNN) using Cloud

International Journal of Recent Technology and Engineering - 2 ◽

10.35940/ijrte.b1041.0982s1119 ◽

2019 ◽

Vol 8 (2S11) ◽

pp. 249-254 ◽

Cited By ~ 1

Keyword(s):

Neural Network ◽

Preterm Birth ◽

Multilayer Perceptron ◽

Zika Virus ◽

Early Prediction ◽

Prediction System ◽

Neural Network Classifier ◽

The World ◽

Accuracy Level ◽

Time Of Infection

Zika virus a mosquito borne flavivirus disease, which is spreading hastily across all over the world. Nearly 95 countries are infected with Zika, Aedes aegypti Mosquitoes is the source of spreading the virus. Microcephaly, myelitis, Guillain – Barre Syndrome and neuropathy are the causes of ZVD. Miscarriages and preterm birth also possible also occur during the time of infection. To overcome an early prediction system is used for detecting the virus using symptoms. The zika dataset is stored in cloud and in our proposed work a Multilayer Perceptron Neural Network classifier used for predicting the Zika virus. The classifier produces accuracy level of 97% the highest accuracy level. Based on the symptoms ZVD is predicted at earlier stage, if they found as infected RNA test will be taken for the concerned person.

Download Full-text