PRESENTATION OF THE TEXT INFORMATION FOR THE ANALYSIS OF TEXT TONALITY BY ARTIFICIAL NEURAL NETWORK. THE IMPLEMENTATION OF PRIVATE DICTONARY METHOD

Abstract

To train the artificial neural network to analyze the text, you should submit input data in a numeric expression. This paper proposes a method for isolating and obtaining numerical characteristics of the text for tonality evaluation. The resulting characteristics in vector form can be transmitted to machine learning algorithm, for determining the classification of texts and tonality. This article contains a ready verbal algorithm for implementation on high-level programming languages. This material is a preparatory stage in the construction of an artiicial neural network, which addressed the issue of representation and normalization of data.