Spelling error correction
based on phonetics combines with context-specific retrieval would eliminate the
unnecessary list of words that are offered but not relevant to the user. In addition,
ontology-based search can be utilized for searching words specific to the domain
thus eliminating words that are not within the specific domain. In this section we
consider a few approaches for dealing with phonetic errors in medical contexts,
which have proven to be the most frequently encountered misspellings identified
from our prior usability study of medical portals. We propose a system architecture,
which is based on these approaches. The main advantage of this system is the ability
to recognize terms relevant to the users??™ context as well as provide definitions
for these terms.
The following section describes the phonetic and semantic-based search for such
a system.
Damerau??™s Simple Detection Method
By using Damerau??™s (1964) simple method we can see that 80% of all spelling errors
are the results of the following four methods. Damerau??™s method was developed
for English but it can be applied in medical terms as in the following example.
Pages:
778
779
780
781
782
783
784
785
786
787
788
789
790
791
792
793
794
795
796
797
798
799
800
801
802