23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
|
# File 'lib/evoc_cli/info.rb', line 23
def dict
$stdout.puts "
keyword description
------- -----------
case_id: user provided tag for the history used
granularity: granularity of the history used
scenario_id: a unique indentifier for this scenario
tx_id: the sha of the commit that the query was sampled from
tx_index: the index of this transaction in the used history (0 is oldest)
tx_size: the number of items in the transaction
query_size: the number of items in the query
query_percentage: query_size/tx_size
expected_outcome_size: |tx - query|
model_size: number of previous transactions relative to this one
model_hours: time span from the first transaction to this one
model_age: number of transactions between end of model and this transaction
max_size: transactions larger than this are filtered out before generating rules
filtered_model_size: model size after the max_size filtering
algorithm: the mining algorithm used to generate the recommendation
aggregator: the aggregation function used to aggregate the rules of the recommendation
measures: the interestingnessmeasures used to rank each rule
recommendation_tag: a unique identifiter of the rules used as a basis for the recommendation
time_rulegeneration: how long it took to generate the rules
time_measurecalculation: how long it took to calculate the measures for each rule
time_aggregation: how long it took to aggregate the rules
number_of_baserules: number of rules before aggregation
number_of_rules: number of rules after aggregation (equal to number_of_baserules when not aggregating)
number_of_hyperrules: number of hyper rules after aggregating
mean_hyper_coefficient: average number of rules aggregated in each hyper rule
largest_antecedent: number of items in the largest antecedent (lhs of rule)
t_ap: average precision where ties are accounted for
ap: the average precision
precision: ratio of correct to incorrect items
precision10: ratio of correct to incorrect items in the top 10
recall: ratio of correct items in recommendation to full set of expected items
recall19: ratio of correct items in recommendation to full set of expected items in the top 10
mean_confidence: the average confidence of the rules in this recommendation
discernibility: the number of uniquely weighted rules to the number of rules
applicable: 1 if rules were generated, 0 otherwise
f1: the f1 measure
first_relevant: the rank of the first correct item
last_relevant: the rank of the last correct item
"
end
|