eps 0.3.0 → 0.3.1
This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
- checksums.yaml +4 -4
- data/CHANGELOG.md +12 -5
- data/README.md +34 -0
- data/lib/eps.rb +19 -10
- data/lib/eps/base_estimator.rb +35 -129
- data/lib/eps/data_frame.rb +7 -1
- data/lib/eps/evaluators/linear_regression.rb +1 -1
- data/lib/eps/label_encoder.rb +7 -3
- data/lib/eps/lightgbm.rb +36 -76
- data/lib/eps/linear_regression.rb +26 -79
- data/lib/eps/metrics.rb +24 -12
- data/lib/eps/model.rb +6 -6
- data/lib/eps/naive_bayes.rb +2 -139
- data/lib/eps/pmml.rb +14 -0
- data/lib/eps/pmml/generator.rb +422 -0
- data/lib/eps/pmml/loader.rb +241 -0
- data/lib/eps/version.rb +1 -1
- metadata +7 -5
- data/lib/eps/pmml_generators/lightgbm.rb +0 -187
data/lib/eps/version.rb
CHANGED
metadata
CHANGED
|
@@ -1,14 +1,14 @@
|
|
|
1
1
|
--- !ruby/object:Gem::Specification
|
|
2
2
|
name: eps
|
|
3
3
|
version: !ruby/object:Gem::Version
|
|
4
|
-
version: 0.3.
|
|
4
|
+
version: 0.3.1
|
|
5
5
|
platform: ruby
|
|
6
6
|
authors:
|
|
7
7
|
- Andrew Kane
|
|
8
8
|
autorequire:
|
|
9
9
|
bindir: bin
|
|
10
10
|
cert_chain: []
|
|
11
|
-
date: 2019-
|
|
11
|
+
date: 2019-12-06 00:00:00.000000000 Z
|
|
12
12
|
dependencies:
|
|
13
13
|
- !ruby/object:Gem::Dependency
|
|
14
14
|
name: lightgbm
|
|
@@ -16,14 +16,14 @@ dependencies:
|
|
|
16
16
|
requirements:
|
|
17
17
|
- - ">="
|
|
18
18
|
- !ruby/object:Gem::Version
|
|
19
|
-
version: 0.1.
|
|
19
|
+
version: 0.1.7
|
|
20
20
|
type: :runtime
|
|
21
21
|
prerelease: false
|
|
22
22
|
version_requirements: !ruby/object:Gem::Requirement
|
|
23
23
|
requirements:
|
|
24
24
|
- - ">="
|
|
25
25
|
- !ruby/object:Gem::Version
|
|
26
|
-
version: 0.1.
|
|
26
|
+
version: 0.1.7
|
|
27
27
|
- !ruby/object:Gem::Dependency
|
|
28
28
|
name: nokogiri
|
|
29
29
|
requirement: !ruby/object:Gem::Requirement
|
|
@@ -117,7 +117,9 @@ files:
|
|
|
117
117
|
- lib/eps/metrics.rb
|
|
118
118
|
- lib/eps/model.rb
|
|
119
119
|
- lib/eps/naive_bayes.rb
|
|
120
|
-
- lib/eps/
|
|
120
|
+
- lib/eps/pmml.rb
|
|
121
|
+
- lib/eps/pmml/generator.rb
|
|
122
|
+
- lib/eps/pmml/loader.rb
|
|
121
123
|
- lib/eps/statistics.rb
|
|
122
124
|
- lib/eps/text_encoder.rb
|
|
123
125
|
- lib/eps/utils.rb
|
|
@@ -1,187 +0,0 @@
|
|
|
1
|
-
module Eps
|
|
2
|
-
module PmmlGenerators
|
|
3
|
-
module LightGBM
|
|
4
|
-
private
|
|
5
|
-
|
|
6
|
-
def generate_pmml
|
|
7
|
-
feature_importance = @feature_importance
|
|
8
|
-
|
|
9
|
-
data_fields = {}
|
|
10
|
-
data_fields[@target] = @labels if @labels
|
|
11
|
-
@features.each_with_index do |(k, type), i|
|
|
12
|
-
# TODO remove zero importance features
|
|
13
|
-
if type == "categorical"
|
|
14
|
-
data_fields[k] = @label_encoders[k].labels.keys
|
|
15
|
-
else
|
|
16
|
-
data_fields[k] = nil
|
|
17
|
-
end
|
|
18
|
-
end
|
|
19
|
-
|
|
20
|
-
build_pmml(data_fields) do |xml|
|
|
21
|
-
function_name = @objective == "regression" ? "regression" : "classification"
|
|
22
|
-
xml.MiningModel(functionName: function_name, algorithmName: "LightGBM") do
|
|
23
|
-
xml.MiningSchema do
|
|
24
|
-
xml.MiningField(name: @target, usageType: "target")
|
|
25
|
-
@features.keys.each_with_index do |k, i|
|
|
26
|
-
# next if feature_importance[i] == 0
|
|
27
|
-
# TODO add importance, but need to handle text features
|
|
28
|
-
xml.MiningField(name: k) #, importance: feature_importance[i].to_f, missingValueTreatment: "asIs")
|
|
29
|
-
end
|
|
30
|
-
end
|
|
31
|
-
pmml_local_transformations(xml)
|
|
32
|
-
|
|
33
|
-
case @objective
|
|
34
|
-
when "regression"
|
|
35
|
-
xml_segmentation(xml, @trees)
|
|
36
|
-
when "binary"
|
|
37
|
-
xml.Segmentation(multipleModelMethod: "modelChain") do
|
|
38
|
-
xml.Segment(id: 1) do
|
|
39
|
-
xml.True
|
|
40
|
-
xml.MiningModel(functionName: "regression") do
|
|
41
|
-
xml.MiningSchema do
|
|
42
|
-
@features.each do |k, _|
|
|
43
|
-
xml.MiningField(name: k)
|
|
44
|
-
end
|
|
45
|
-
end
|
|
46
|
-
xml.Output do
|
|
47
|
-
xml.OutputField(name: "lgbmValue", optype: "continuous", dataType: "double", feature: "predictedValue", isFinalResult: false) do
|
|
48
|
-
xml.Apply(function: "/") do
|
|
49
|
-
xml.Constant(dataType: "double") do
|
|
50
|
-
1.0
|
|
51
|
-
end
|
|
52
|
-
xml.Apply(function: "+") do
|
|
53
|
-
xml.Constant(dataType: "double") do
|
|
54
|
-
1.0
|
|
55
|
-
end
|
|
56
|
-
xml.Apply(function: "exp") do
|
|
57
|
-
xml.Apply(function: "*") do
|
|
58
|
-
xml.Constant(dataType: "double") do
|
|
59
|
-
-1.0
|
|
60
|
-
end
|
|
61
|
-
xml.FieldRef(field: "lgbmValue")
|
|
62
|
-
end
|
|
63
|
-
end
|
|
64
|
-
end
|
|
65
|
-
end
|
|
66
|
-
end
|
|
67
|
-
end
|
|
68
|
-
xml_segmentation(xml, @trees)
|
|
69
|
-
end
|
|
70
|
-
end
|
|
71
|
-
xml.Segment(id: 2) do
|
|
72
|
-
xml.True
|
|
73
|
-
xml.RegressionModel(functionName: "classification", normalizationMethod: "none") do
|
|
74
|
-
xml.MiningSchema do
|
|
75
|
-
xml.MiningField(name: @target, usageType: "target")
|
|
76
|
-
xml.MiningField(name: "transformedLgbmValue")
|
|
77
|
-
end
|
|
78
|
-
xml.Output do
|
|
79
|
-
@labels.each do |label|
|
|
80
|
-
xml.OutputField(name: "probability(#{label})", optype: "continuous", dataType: "double", feature: "probability", value: label)
|
|
81
|
-
end
|
|
82
|
-
end
|
|
83
|
-
xml.RegressionTable(intercept: 0.0, targetCategory: @labels.last) do
|
|
84
|
-
xml.NumericPredictor(name: "transformedLgbmValue", coefficient: "1.0")
|
|
85
|
-
end
|
|
86
|
-
xml.RegressionTable(intercept: 0.0, targetCategory: @labels.first)
|
|
87
|
-
end
|
|
88
|
-
end
|
|
89
|
-
end
|
|
90
|
-
else # multiclass
|
|
91
|
-
xml.Segmentation(multipleModelMethod: "modelChain") do
|
|
92
|
-
n = @trees.size / @labels.size
|
|
93
|
-
@trees.each_slice(n).each_with_index do |trees, idx|
|
|
94
|
-
xml.Segment(id: idx + 1) do
|
|
95
|
-
xml.True
|
|
96
|
-
xml.MiningModel(functionName: "regression") do
|
|
97
|
-
xml.MiningSchema do
|
|
98
|
-
@features.each do |k, _|
|
|
99
|
-
xml.MiningField(name: k)
|
|
100
|
-
end
|
|
101
|
-
end
|
|
102
|
-
xml.Output do
|
|
103
|
-
xml.OutputField(name: "lgbmValue(#{@labels[idx]})", optype: "continuous", dataType: "double", feature: "predictedValue", isFinalResult: false)
|
|
104
|
-
end
|
|
105
|
-
xml_segmentation(xml, trees)
|
|
106
|
-
end
|
|
107
|
-
end
|
|
108
|
-
end
|
|
109
|
-
xml.Segment(id: @labels.size + 1) do
|
|
110
|
-
xml.True
|
|
111
|
-
xml.RegressionModel(functionName: "classification", normalizationMethod: "softmax") do
|
|
112
|
-
xml.MiningSchema do
|
|
113
|
-
xml.MiningField(name: @target, usageType: "target")
|
|
114
|
-
@labels.each do |label|
|
|
115
|
-
xml.MiningField(name: "lgbmValue(#{label})")
|
|
116
|
-
end
|
|
117
|
-
end
|
|
118
|
-
xml.Output do
|
|
119
|
-
@labels.each do |label|
|
|
120
|
-
xml.OutputField(name: "probability(#{label})", optype: "continuous", dataType: "double", feature: "probability", value: label)
|
|
121
|
-
end
|
|
122
|
-
end
|
|
123
|
-
@labels.each do |label|
|
|
124
|
-
xml.RegressionTable(intercept: 0.0, targetCategory: label) do
|
|
125
|
-
xml.NumericPredictor(name: "lgbmValue(#{label})", coefficient: "1.0")
|
|
126
|
-
end
|
|
127
|
-
end
|
|
128
|
-
end
|
|
129
|
-
end
|
|
130
|
-
end
|
|
131
|
-
end
|
|
132
|
-
end
|
|
133
|
-
end
|
|
134
|
-
end
|
|
135
|
-
|
|
136
|
-
def xml_segmentation(xml, trees)
|
|
137
|
-
xml.Segmentation(multipleModelMethod: "sum") do
|
|
138
|
-
trees.each_with_index do |node, i|
|
|
139
|
-
xml.Segment(id: i + 1) do
|
|
140
|
-
xml.True
|
|
141
|
-
xml.TreeModel(functionName: "regression", missingValueStrategy: "none", noTrueChildStrategy: "returnLastPrediction", splitCharacteristic: "multiSplit") do
|
|
142
|
-
xml.MiningSchema do
|
|
143
|
-
node_fields(node).uniq.each do |k|
|
|
144
|
-
xml.MiningField(name: display_field(k))
|
|
145
|
-
end
|
|
146
|
-
end
|
|
147
|
-
node_pmml(node, xml)
|
|
148
|
-
end
|
|
149
|
-
end
|
|
150
|
-
end
|
|
151
|
-
end
|
|
152
|
-
end
|
|
153
|
-
|
|
154
|
-
def node_fields(node)
|
|
155
|
-
fields = []
|
|
156
|
-
fields << node.field if node.predicate
|
|
157
|
-
node.children.each do |n|
|
|
158
|
-
fields.concat(node_fields(n))
|
|
159
|
-
end
|
|
160
|
-
fields
|
|
161
|
-
end
|
|
162
|
-
|
|
163
|
-
def node_pmml(node, xml)
|
|
164
|
-
xml.Node(score: node.score) do
|
|
165
|
-
if node.predicate.nil?
|
|
166
|
-
xml.True
|
|
167
|
-
elsif node.operator == "in"
|
|
168
|
-
xml.SimpleSetPredicate(field: display_field(node.field), booleanOperator: "isIn") do
|
|
169
|
-
xml.Array(type: "string") do
|
|
170
|
-
xml.text node.value.map { |v| escape_element(v) }.join(" ")
|
|
171
|
-
end
|
|
172
|
-
end
|
|
173
|
-
else
|
|
174
|
-
xml.SimplePredicate(field: display_field(node.field), operator: node.operator, value: node.value)
|
|
175
|
-
end
|
|
176
|
-
node.children.each do |n|
|
|
177
|
-
node_pmml(n, xml)
|
|
178
|
-
end
|
|
179
|
-
end
|
|
180
|
-
end
|
|
181
|
-
|
|
182
|
-
def escape_element(v)
|
|
183
|
-
"\"#{v.gsub("\"", "\\\"")}\""
|
|
184
|
-
end
|
|
185
|
-
end
|
|
186
|
-
end
|
|
187
|
-
end
|