Information system for the intellectual assessment customers text reviews tonality based on artificial neural networks

Abstract

This article presents the results of the concept development and software information system for assessing text data tonality implementation by users based on artificial neural networks. The main problems in this topic are identified, the features of using deep machine learning for the text data mining problems are presented. An information system project has been developed, the preprocessing procedure and data filtering algorithms have been described, the specifics of data normalization for formalizing artificial neural network models are formalized. The options for using the information system, the block structure, the interface prototype and the procedure for user interaction with the software application are developed. The training effectiveness study results and the use of an artificial neural network model to solve the tasks are presented, the most suitable values of hyperparameters that have a primary impact on the model quality are identified and selected.