Short title: TU-Mars

Long title: Text Understandability for Maintenance, Availability, Reliability and Safety Management Systems

Sponsor: TU Wien

Assignment to our (IMW) research priorities:

  • Identification of research gaps in the field of knowledge extraction from text in the context of industrial maintenance
  • Identification of features contributing to an automated human-like text comprehension
  • Application of text mining and machine learning techniques, in order to extract valuable text understandability features from text
  • Identification of impact of text understandability in industrial maintenance

Duration: 2020-ongoing

TU-MARS: Text Understandability in Industrial Maintenance in form of a graphic

Abstract:

Textual data majorly reflects objective and subjective human specific knowledge. Focusing on big data in industrial and operation management, the value of textual data is oftentimes undermined. The scientific challenge is to effectively discover knowledge from text data and convert it into automated processes for inferential reasoning, predicting and prescribing. TU-MARS, a proof-of-concept software demonstrator, employs text mining techniques and machine learning algorithms to evaluate the text understandability, based on various measures including text readability, association measurement and sentiment analysis, of real-world manufacturing datasets. The proposed text understandability enables an early stage detection of failure, the reduction of human failures and leads to an immense improvement of explication of human knowledge.

Results:

The project, implements a proof-of-concept software demonstrator (TU-MARS), that offers:

  • Text understandability dashboard for each analyzed report
  • Extraction of text understandability features on report level and on feature level
  • Presentation of detailed text readability measures for each analyzed report
  • Presentation of extracted sentiment for each analyzed report
  • Presentation of extracted associations and an association measurement index for each analyzed report

Partners: TU Wien, Research Unit Production and Maintenance Management

Contact details:

Dipl.-Ing. Theresa Madreiter

Telephone: +43 1 58801 33051

Email: theresa.madreiter@tuwien.ac.at