UNSTRUCTURED DATA EXTRACTION AND INTELLIGENT DOCUMENT PROCESSING

Background. This paper examines modern approaches to data structuring and information enrichment in the context of digitalization. Approaches to solving the problem of automated data extraction are proposed using the example of tool manufacturers' catalogs in order to minimize manual labor, reduce time costs and apply them in the «TKMP.Istok» marketplace. Materials and methods. Methods for detecting objects using the YOLOv10 model and optical character recognition using the EasyOCR and PaddleOCR models are considered. Results and conclusions. As a result of the experimental application, the processing time of documents has been reduced by more than 90 %. The methods used make it possible to obtain high accuracy of information extraction, which opens up new prospects for the use of artificial intelligence in business processes as a key component of digital transformation.