THIS IS THE CONJECTURE-LIST STAGE (4), which requires first the completion of stage 1 with File A replaced by File AA, a broad-category hypothesis (such as deficiencies, diet, toxins, or drugs) instead of a specific agent. From the broad-category AA-titles, Stage 4 constructs a list of individual terms, A, within those titles, and ranks them according to the number of different bridging terms, B. The A-list can be edited either by deleting terms, or by grouping terms. Grouping is partly automated by pre-grouping together phrases that have a common keyword, the latter shown as a header. Headers form the basis for ranking. Clicking on any A-term header will display the associated B-terms and the corresponding A and C titles.

If the resulting A-list seems unmanageably large, go back to Stage 2 and delete unwanted terms from the original (AA-based) B-list.

If you have already completed Stage 4, but wish to form new groups or delete more A-terms, then go to STAGE 5 XU instead. If you just want to browse previous results, go to STAGE 3 XU

CAUTION!: If you have already edited the A-list, and wish to save and/or continue your editing, go to Stage 5; DO NOT RE-ENTER STAGE 4. Stage 4 starts over, ignoring your previous editing (but does not influence the original AA-based B-list).

(Note that this works differently from Stage 2 (B-list editor), which picks up from where you left off each time you enter it). To proceed with STAGE 4, enter email address or user code:

This stage implements a simplified version of Procedure I, described in Artificial Intelligence 91 (1997) 183-203.

MAIN INPUT (STAGE 1)
B-EDIT (STAGE 2)
View Results (STAGE 3)
Continue A-EDIT (STAGE 5)
ARROWSMITH HOME PAGE.