A Method of Text Sample Size Adaptation Based on Ontological and Cognitive Analysis

In this paper the issue of the latency of the cognitive assistant system, which functions on the basis of the stylometry, is considered. The conducted research allows concluding that the system latency can be decreased by means of the text sample size adaptation. The first approach to this is to correlate the text samples with the data source types and the methods of the text feature vector forming. Besides, the users’ feedbacks on the cognitive assistant functioning quality can be involved into the text sample size determination. In this paper, the new method of text sample size adaptation is proposed and discussed. It includes two approaches: the first one is the ontologycal analysis based on the data source type attribute, and the second one is the cognitive analysis of the users’ feedbacks. The mixture of these approaches allows to decrease the text sample size and so to decrease the time needed for the text feature vector forming.