COMMENT ⊗   VALID 00002 PAGES
C REC  PAGE   DESCRIPTION
C00001 00001
C00002 00002	.E
C00006 ENDMK
C⊗;
.E

\8}\7{\it LEVEL}\→{\it e.g.} \→{\it # Con's}\→{\it # w/Heur}\→{\it # Heurs}\→{\it Avg}\8 \7
\ {\it Avg w/Heur}\8 \7{\it # Fillin}\→{\it # Sugg}\→{\it # Check}\→{\it # Int}
\8}\3

\8}\3
→0\→Anything\→1\→1\→10\→10.0\→10.0\→0\→5\→0\→5
\8}\3

\8}\3
→1\→Any-Concept\→1\→1\→110\→110.0\→110.0\→39\→30\→20\→21
\8}\3

\8}\3
→2\→Active\→2\→2\→24\→12.0\→12.0\→7\→10\→4\→3
\8}\3

\8}\3
→3\→Operation\→6\→3\→31\→5.2\→10.3\→11\→3\→3\→14
\8}\3

\8}\3
→≥4\→Union\→100\→11\→63\→0.6\→5.7\→26\→15\→8\→16
\8}\3

.TURN OFF ``α``; TURN ON ``∞→``; SELECT 8;

%∞α→$

.ESS

Here is a key to the column headings:

.BN; INDENT 4,8,0; PREFACE 0; SPACING 0;

LEVEL: How far down the Genl/Spec tree of concepts we are looking.

e.g.: A sample concept at that level.

# Con's: The total number of concepts at that level.

# w/Heur: How many of them have \4some\0 heuristics.

# Heurs: The total number of heuristics attached to concepts at that level.

Avg: (# Heurs) / (# Concepts); i.e., the mean number of heuristics per concept,
at that level.

Avg w/Heur: (# Heurs) / (# w. Heurs)

# Fillin: Total number of ``Fillin" type heuristics at that level.

# Sugg: Total number of ``Suggest" type heuristics at that level.

# Check: Total number of ``Check" type heuristics at that level.

# Int: Total number of ``Interestingness" type heuristics at that level.

.ESS


The heuristic rules are seen \4not\0 to be distributed uniformly, homogeneously
among all the initial concepts. The extent of this skewing was not realized
by the author until the above table was constructed.
A surprising proportion of rules are attached to the very general concepts.
The top 10% of the concepts contain 73% of all the heuristics.
One notable exception is the ``Interest" type heuristics: they seem more evenly
distributed throughout the tree of initial concepts.
This tends to suggest that future work on providing ``meta-heuristics" should
concentrate on how to automatically synthesize those Interest heuristics for
newly-created concepts.