Differences

This shows you the differences between two versions of the page.

--- ivci:precise [2024/01/05 10:43] – [Possible measure] fkaag
+++ ivci:precise [2024/04/02 15:38] (current) – fkaag
@@ Line 9: / Line 9: @@
 Precision ensures individuals receive the necessary vaccines for full protection.
-===== Possible measure =====
+===== Measure =====
-Code systems including concrete codes are unambiguous. If the product can be identified, it can be codified. For code systems containing only abstract codes, we estimate the level of ambiguity as the average number of concrete vaccines within a class (evaluated against the codes for concrete vaccines present in NUVA ). Unambiguousness is computed as 100 divided by this number.
+If we consider [[NUVA]] as the most comprehensive list of possible vaccine records, precision is characterized by the ability to distinguish between NUVA concepts. For each NUVA concept (within the codeable ones, see [[complete|completeness]]), there is at least one code in the code system that matches with a minimum set of NUVA concepts. If this set is reduced to 1, then the concept designation is absolutely precise.
->FK - If we consider NUVA as the most comprehensive list of possible vaccine records, precision is characterized by the ability to distinguish between NUVA concepts. For each NUVA concept (within the codeable ones, see [[complete|completeness]]), there is at least one code in the code system that matches with a minimum set of NUVA concepts. If this set is reduced to 1, then the concept designation is absolutely precise. We can then build a distribution of the number of NUVA concepts per size of their most precise designation, and use it to elaborate the precision index for the whole code system as the weighted average of these sizes.
+The overall ambiguity of the code system is then characterized as the square root of the average size of these sets for all NUVA codes. The square root is used because the size of the set impacts twice the average. With this method, if all sets of indiscriminate concepts had the same size, the ambiguity index would be equal to that size.
+The precision is the inverse of the ambiguity.
+The Python program described at [[nuva-utils]] computes the values for completeness and precision, based upon an alignment table between a given code system and NUVA codes. The description pages for each code (see Code systems) includes these values as computed from the currently known alignments.