Methodology for knowledge extraction for unstructured data by means of onto the voltage-to-current based text - mining

摘要

It is a method for extracting data from service repair - report word are provided in a car inspection service - a reporting system. Each service repair - report word according to, the comments of a technician relating to a component, a symptom associated with the component and a the symptom associated repair measure. Each service repair - report word according to includes information relating to an identified problem in connection with at least one vehicle part. It is a diagnosis - and - ontology the prognosis of the data bank is provided which, by means of a vehicle component classification, a vehicle part - subclasses - classification and a relationship classification is structured, wherein the relationship classification symptom - relationships and measure - relationships comprises. Each of the service repair - report word phones, with the use of the diagnosis - and - ontology the prognosis of the data bank structured. Combinations of information are made of the structured service repair - report word loud depending on user - input criteria extracted. For each extracted combination, a frequency is determined, with which the latter in the structured service repair - report word loud occurs. The service repair - report word sounds, for each combination of in each case grouped to form a cluster.