Level: Intermediate
There are noteworthy examples of text mining techniques being used in engineering departments as well, such as analyzing quality trends and reporting on optimized maintenance through text mining of maintenance records. This report introduces a case where quality and productivity problems were discovered and their causes were identified by text mining production performance data.
The production performance data contains dozens of variables, such as machine and operator numbers, processing conditions, material conditions, and special notes. Special notes are text data manually entered by operators when irregular events occur, and because the content includes words unrelated to abnormal events, these are registered as stop words and excluded from analysis. Next, synonyms such as "mistake, mistake, mistake" are replaced with a single word through recoding. After this preprocessing, the presence or absence of frequently occurring abnormal words is saved in a column, and by analyzing this presence or absence in relation to two variables, such as production conditions, correlations with abnormalities such as machine, operator, processing conditions, and material conditions became clear, enabling the cause of the abnormality to be determined.
