UDC 004.912:656.135
INTEGRATION OF NLP TECHNOLOGIES AND PREDICTIVE MODELS FOR CARGO TRANSPORTATION MARKET ANALYSIS IN ORGANIZATIONAL SYSTEMS
D. V. Bokarev, postgraduate student; Russian Academy of National Economy and Public Administration
under the President of the Russian Federation; Moscow, Russia;
orcid.org/0009-0005-9883-9772, e-mail: This email address is being protected from spambots. You need JavaScript enabled to view it.
The article is devoted to the development and testing of the architecture for the integration of natural
language processing (NLP) technologies and predictive machine learning models for analyzing the trucking
market in organizational systems of digital platforms. The relevance of the research is due to the fact that
modern freight transportation market is characterized by a significant amount of unstructured data (dialogues,
correspondence, voice messages) containing valuable information about demand, routes and prices
that are not taken into account by traditional analytical tools based mainly on registers of completed transactions. This limits the accuracy of forecasting and the effectiveness of management decisions. The aim of
the work is to develop and experimentally test end-to-end architecture combining NLP module for extracting
transaction attributes with XGBoost predictive model for analyzing price dynamics and demand. Among the
tasks: the formation of a corpus of dialogues, further training of RuBERT model for extracting entities, the
construction of feature space, training and validation of a predictive model as well as assessment of management effects of implementation. The empirical base was formed by dialogues of digital platform. The accuracy of NER extraction reached F1 = 0,887. The forecast model demonstrated a MAPE of 5,04 % for urban
transportation (1 day horizon). The inclusion of NLP data reduced the forecast error by 1,34 percentage
points and increased R2 to 0,90. The implementation of the system provided management effects: a 1,1 percentage point increase in conversion. It is concluded that the integration of NLP and predictive analytics
forms a new competitive advantage of platforms based on the transformation of unstructured communications
into a management forecasting tool.
Key words: NLP, price forecasting, organizational systems, management, freight transportation, digital
platforms, XGBoost, data extraction, transportation services, competitiveness