Development of Scenarios for Automatic Processing and Data Mining in a Multi-Agent Environment

Abstract

The article proposes tools for the creation and execution of scenarios processing and data mining in the multi-agent environment. The hybrid database of information resources is developed, the database of methods and models, ontologies and tools for processing and extracting data in the web environment is also proposed. Metadata describes machine learning methods related to regression analysis and classification of information resources based on Bayesian theorem, decision trees, support vector machines, convolutional neural networks, and other. Examples of research scenarios, including the processing of video images and signals, the classification and identification of models of object trajectories are given. The task of automated data processing and analysis is considered with use of the example of multidimensional sonar monitoring. Automatic data analysis is implemented within the network software environment based on Semantic Web standards. The results of the work can be applied in research and in the field of students training.