Shrestha S. Quantitative estimations of phonosemantics characteristics of a natural language texts and their use in artificial intelligence systems

The PhD work is devoted to questions of research of emotional perception by the person of the textual information. The sound-letter has been chosen as entrance natural language unit. The realization strategy of the multilevel linguistic experiment is offered on the basis of the generated pseudo-words which is insensible of obligations and as much as possible approached under characteristics to the Russian text. The algorithm for iterative getting of sound-letter location coefficients is developed with usage of a method of the genetic algorithm. These coefficients are describing influence on the person an emotional perception of the textual information. The deduced values of coefficients considerably are conductive to more adequate modeling of the emotional text estimation. The constructed model of an emotional perception estimation of the textual information is included in the system of phonosemantic examination of the text for an estimation of perception of the textual information by comparison numerical estimation subjective.