Unstructured data treatment for big data solutions

We constructed a system infrastructure capable of processing unstructured data, with the aim of practical application of the system for document data analysis in the manufacturing industry. Using past ISSM research paper data, papers were classified and verified. Using morphological analysis, the extracted parts of speech were used as feature quantities, and machine learning was executed. Since effective data was obtained with the paper classification, using the same analysis method, actual manufacturing data logs were analyzed, defect determination was made, and a high accuracy rate of 75% was achieved.