Vysotska O. Methods of biometric authentication of information systems users by their keystroke pattern and handwriting

The work first analyzes the existing types AISU; the classification of biometric recognition systems has been improved; selected biometric characteristics, optimal for authentication, with their help, users of information systems, namely keystroke pattern and handwriting; a comparative analysis of the most common recognition methods by selected biometric characteristics was performed. Existing types of neural networks were analyzed and their most suitable variant was chosen for solving the AISU problem, namely probabilistic neural network (PNN). The methods AISU by their keystroke pattern and handwriting by means of PNN are offered and programmatically implemented. An important component of advanced methods is developed methods of primary processing of samples of keystroke patterns and handwriting, using these methods increases the probability of correct recognition of in-formation systems users.