Page 231 - Data Science Algorithms in a Week

Managing Overcrowding in Healthcare using Fuzzy Logic 215

Table 7: Results of expert evaluation for subsystem I’s fuzzy rules.

Table 8 is comprised of results from the assessments of the fuzzy rules from
subsystem II. This table shares similar features from Table 7, consisting of the same
number of columns and expert evaluations. Below the table is a legend comprising three
linguistic classes which are color-coded. Within subsystem II, five of the nine rules
received 90% consensus or greater, consisting of FLS2-01, FLS2-04, FLS2-05, FLS2-06,
and FSL2-09. Three of these rules received 80% consensus, which were FLS2-02, FSL2-
07, and FSL2-08. FSL2-03 received 50% consensus. The average consensus rate for the
whole subsystem was 84%, which is higher than the previous subsystem, which featured
more fuzzy rules and linguistic classes. Seven of the evaluated fuzzy rules were assessed
with only two linguistic terms or less, and two rules were assessed with three terms. The
frequency of linguistic classes used in assessing rules was the highest in “inadequate”
with 41 uses, followed by “partially adequate”, and “adequate”, with 34 and 15 uses,
respectively.

Table 8: Results of expert evaluation for subsystem II’s fuzzy rules.

The final fuzzy rule statements for subsystem II (Appendix B) after consensus are
listed according to their rule number. These final nine rules are stored in the fuzzy rule
base of subsystem II to feed the decision engine of the fuzzy system.
Table 9 contains data from the expert assessments of the fuzzy rules of subsystem III.
It is structured in the same manner as the previous fuzzy rule evaluation tables in terms of
the number of columns and what they represent, however there are four color-coded