Using Artificial Neural Networks in Predicting the Level of Stress among Military Conscripts

The present study aims to elucidate the main variables that increase level of stress at beginning military conscription service using an artificial neural network (ANN)-based prediction model. Random sample data were obtained from one battalion Lithuanian Armed Forces, and a survey was conducted generate for training testing ANN models. Using nonlinearity in research, numerous structures constructed verified limit optimal number neurons, hidden layers, transfer functions. highest accuracy by multilayer perceptron (MLPNN) with 6-2-2 partition. A standardized rescaling method used covariates. For activation function, hyperbolic tangent 20 units layer as well back-propagation algorithm. best model determined showed smallest cross-entropy error, correct classification rate, area under ROC curve. These findings show, high precision, cohesion team adaptation routines are two critical elements have greatest impact on conscripts.