Neural network natural language processing tools for identifying personal priorities in the project performers selection in the field of smart agriculture

Abstract

Abstract The project implementation effectiveness in creating digital smart agriculture systems depends on the correct selection of performers. Considering personal priorities makes it possible to increase the validity of decision-making regarding the employment of specific individuals for the implementation of IT projects in the agro-industrial sector. Personal priorities are internal, hidden characteristics that have an effect on the process of long-term joint work and interaction in various situations that arise in the team. The identification of the analyzed individual’s personal priorities is proposed to be reduced to solving the classification problem based on the analysis of person’s text Internet traces using neural network technologies of natural language processing. As a training sample, it is proposed to use a set of text document vectors and the corresponding marks of personal priority classes. In the process of identifying the personal priorities classes, it is required to create an appropriate text array based on parsing and processing of text messages published on the Internet by the analyzed person. Next, from the resulting text array, a text vector of the document must be formed, which then needs to be submitted to the input of the neural network. It is assumed that the mark of the analyzed individual’s personal priority class will be displayed in the output layer of neurons.