Jdi tool uses statistical associations between words and JDs, between MHs and JDs, and between SHs and JDs, from a training set of MEDLINE citations. The word-Jd scores, Mh-Jd scores, and Sh-Jd scores are pre-calculated and loaded into a database. Jdi takes the inputs, which may be text phrases, MeSH terms, or a combination. Filters are applied to text input, such as word extraction algorithms, stopwords, minimum word length, etc. Then, JDI calculates the average score for all inputs, and sends the ranked JDs with their scores to the output.
Jdi is the core methodology of TC tools. It is used in Sti and Stri program. It is used to categorize text, index contents, retrieve records, and Word Sense Disambiguation (more).