Dissertation Information for Diane Vizine-GoetzNAME:
DEGREE:
DISCIPLINE:
SCHOOL: ADVISORS: COMMITTEE MEMBERS: MPACT Status: Fully Complete Title: A Computer Algorithm for Correcting Spelling and Typographical Errors in Subject Headings
Abstract: A technique for correcting spelling and typographical errors in Library of Congress subject headings occurring in tropical and geographic subject headings occurring in tropical and geographic subject heading fields of bibliographic records in the OCLC data base is tested. A three phase error correction procedure is employed. The three major steps consist of : 1) creation of an initial subject heading file, 2) formation of subject heading keys, and 3) correction of spelling and typographical errors. The technique compares each entry under consideration for possible errors to every other entry which: (1) has the same initial characters, (2) has a valid status indicator, (3) and is within one character in length of the entry being checked. The entry being checked is identified as an addition error if its key length is one more than that of the key length of the entry to which it is being compared. if the key length is one less, it is is identified as an omission error. If the key lengths are equal and one character does not match, the entry is assumed to be a substitution error. A transposition error is identified when the key lengths are the same and two adjacent characters do not match. In this way all errors of addition, omission, substitution, and transposition are identified. Corrections were also made for corrections in spacing, punctuation, and capitalization. This study found that, overall, 80% of pairings to identify all occurrences of addition, omission, substitution, and transposition errors were valid. The most common error type corrected was omission. Transposition errors were the least frequently occurring error type. |
MPACT Scores for Diane Vizine-GoetzA = 0 Advisors and Advisees Graphgenerating graph, please reload |
Students under Diane Vizine-Goetz
ADVISEES:
- None
COMMITTEESHIPS:
- None