This function requires the bigdata library section to be loaded.
bd.assoc.rules.graph(rules, column="lift", column.min=NULL, column.max=NULL)
data.frame
or
bdFrame
giving a set of rules, as produced by
.
This must contain either a
"rule"
column,
or individual item columns (
"con1"
,
"ant1"
, etc.),
so the consequent and antecedent items can be extracted for each rule.
rules
containing a numerical value to be plotted for each rule.
column
value
is equal or greater than
column.min
are plotted.
column
value
is equal or less than
column.max
are plotted.
NULL
value is returned.
rules
is produced on the current graphics device.
This plot processes the association rules, collecting a list of all
items that appear as consequents in any rules, and a list of all items
that appear as antecedents in any rules. Each of these lists is
sorted alphabetically and displayed in the graph, with consequent
items displayed in a vertical list along the left side, and the
antecedent items displayed in a list along the bottom side. Then, for
each rule, a symbol is displayed at the intersection of the rule's
consequent item and each of its antecedent items. The symbol is an
unfilled diamond, whose size is proportional to the
column
value for the rule. Because the diamond is not
filled, multiple diamonds may be plotted in the same location, and
still be visible, if they represent rules with different
column
values.
If there are too many rules, the graph may be too crowded. In this
case, the
column.min
and
column.max
values
can be used to only process rules with
column
values in
the specified range.
This plot can be used to get a rough sense of which consequent and
antecedent items appear most often in the rules with high
column
values. Since information from multiple rules may
be plotted over each other, it is not possible to read individual
rules from this graph. To view individual rules, examine the rules
data directly.
bd.assoc.rules.graph( bd.assoc.rules( data.frame(aa=c("A","A","B","B","B"), bb=c("C","C","C","C","D"), stringsAsFactors=F), input.format="item.list") )