Industrial Automatic Control Systems and Controllers Annotation << Back Ways to Raise Accuracy in the Task of Identifying Keyboard Handwriting by a Neural Network Model Mustafaev A.G., Kobzarenko D.N., Shikhsaidov B.I. Based on the source material obtained by taking the keyboard handwriting from users of a personal computer, the task of classifying the user using a neural network is set. Ways to improve recognition accuracy are considered. New class fi ltering algorithms are proposed, based on counting the output data of a neural network to obtain the class number, and analyzing the distribution of types of sequences of a chain of four keyboard events and a method for eliminating some classes based on it. The results of original experiments performed to test hypotheses regarding these methods of increasing accuracy are presented. Keywords: keyboard handwriting, neural network, classifi cation task. Pp. 03-13.