"Intelligent" selections for sampling from raw reliability data
It is proposed that the definition of taxa should be based on collected data on relevant variables outlined by subject matter experts. Then, as not all codes show data close to being equi-distributed and variables are highly dependent on each other, the definition of samples should be carried out by "Intelligent" selection. Selections are made to identify categories of each relevant variable that account for a large fraction of the observations. Data exploration then proceeds with selections against infrequent combinations of the essential categories and the identifications of a relatively small number of taxa where most data lie. This approach was used to carry out a taxonomy for electric motors from CEDB data based on component type, current type, voltage, power, mode of operation, rotational speed, and environmental data on 510 electric motors. This is compared to taxonomies used in published databooks.
Bibliographic Reference: Report: ISEI/IE/2877/95 EN (1995) 10 pp.
Availability: Available from Joint Research Centre, Public Relations Office, I-21020 Ispra (Va.) (IT)
Record Number: 199510908 / Last updated on: 1995-08-10
Original language: en
Available languages: en