Examination of static and 50 Hz electric field effects on tissues by using a hybrid genetic algorithm and neural network

Abstract

The effects of electric fields on tissue are the main subject of many investigations. The importance of this subject comes from the electrical properties of the cell membrane and its sensitivity to changes in electrical conditions. Permeability of membranes to various ions can change by the effect of an electric field depending on their conductivity. The performances of cells and tissues change due to differences between the membrane's permeability to various ions and molecules. The aim of this study was to determine lipid peroxidation and superoxide dismutase (SOD) levels in spleen and testis tissues exposed to different intensities and exposure periods of static and 50 Hz alternating electric fields. The increase in SOD and thiobarbituric acid reactive substance levels of spleen and testis tissues was found to depend significantly on the type of electric field and the exposure period. The experimental results are applied to a hybrid genetic algorithm and neural network as learning data and the training of the feedforward neural network is realized. At the end of this training, without applying electric field to tissues, the determination of the effects of the electric field on tissues by using a computer is predicted by the neural network. After the experiments, the prediction of the hybrid genetic algorithm and neural network approach is on average 99.25%–99.99%.

Design of highly effective multilayer feedforward neural network by using genetic algorithm

“…The backpropagation learning algorithm is the most preferred method for training of the FNNs (Hardalaç & Güler, ; Jackowski & Wozniak, ; Pazera, Buciakowski, Witczak, & Mrugalski, ). Determination of the learning parameters (LPs), which are learning coefficients and momentum coefficients in backpropagation learning algorithm, is another major problem that affects on the performance of the network so the FNN applications have shown that a low value of the LPs decrease learning speed while a high value of the LPs increases the learning speed but may also result in oscillations which lead to no learning at all (Kandil, Khorasani, Patel, & Sood, ).…”

Flood forecasting at the Dadu River in China based on ANN

“…Oh et al (2006) established sub-DFCIs (daily financial condition indicators) for various daily financial variables using an ANN and used the DFCI for the Korean financial market as an empirical case study. Hardalaç and Güler (2008) examined static and 50 Hz electric field effects on tissues and found that the prediction of the hybrid genetic algorithm and neural network approach was on average 99.25-99.99%. In terms of flooding, Campolo et al (1999) developed a neural network model to analyze and forecast the behaviour of the Tagliamento River in Italy during heavy rain periods.…”