Intelligent analysis of short messages for road infrastructure management

Lyapin Artur Mansurovich

In the article analyzes the problems of using methods of data mining by specialists who do not have high qualifications in the field of information technology. This aspect creates difficulties in training classifiers, the selection of trained data sets and their subsequent use, so the development of self-learning methods is an urgent task. Purpose. The purpose of the work is to develop an ensemble of algorithms for intelligent data processing, which allow to classify text messages without prior training of the classifier. Novelty. The model assumes a deep integration of road transport and information technologies on the basis of a single network platform that includes a traffic management system, an alert system for emergency situations, an intelligent system for analyzing geotagged user messages, a Web interface, etc. Practical relevance. The intellectual system represents a mechanism for the classification of short text messages for the timely notification of road users about emergency situations and support for decision-making by service operators

Keywords:data mining, Smart city, k-nearest neighbors, abnormal road situation.