fselector 0.1.0
Sign up to get free protection for your applications and to get access to all the features.
- data/LICENSE +21 -0
- data/README.md +195 -0
- data/lib/fselector.rb +41 -0
- data/lib/fselector/algo_continuous/PMetric.rb +51 -0
- data/lib/fselector/algo_continuous/ReliefF_c.rb +190 -0
- data/lib/fselector/algo_continuous/Relief_c.rb +150 -0
- data/lib/fselector/algo_continuous/TScore.rb +52 -0
- data/lib/fselector/algo_continuous/discretizer.rb +219 -0
- data/lib/fselector/algo_continuous/normalizer.rb +59 -0
- data/lib/fselector/algo_discrete/Accuracy.rb +35 -0
- data/lib/fselector/algo_discrete/AccuracyBalanced.rb +37 -0
- data/lib/fselector/algo_discrete/BiNormalSeparation.rb +45 -0
- data/lib/fselector/algo_discrete/ChiSquaredTest.rb +69 -0
- data/lib/fselector/algo_discrete/CorrelationCoefficient.rb +42 -0
- data/lib/fselector/algo_discrete/DocumentFrequency.rb +36 -0
- data/lib/fselector/algo_discrete/F1Measure.rb +41 -0
- data/lib/fselector/algo_discrete/FishersExactTest.rb +47 -0
- data/lib/fselector/algo_discrete/GMean.rb +37 -0
- data/lib/fselector/algo_discrete/GSSCoefficient.rb +43 -0
- data/lib/fselector/algo_discrete/GiniIndex.rb +44 -0
- data/lib/fselector/algo_discrete/InformationGain.rb +96 -0
- data/lib/fselector/algo_discrete/MatthewsCorrelationCoefficient.rb +45 -0
- data/lib/fselector/algo_discrete/McNemarsTest.rb +57 -0
- data/lib/fselector/algo_discrete/MutualInformation.rb +42 -0
- data/lib/fselector/algo_discrete/OddsRatio.rb +46 -0
- data/lib/fselector/algo_discrete/OddsRatioNumerator.rb +41 -0
- data/lib/fselector/algo_discrete/Power.rb +46 -0
- data/lib/fselector/algo_discrete/Precision.rb +31 -0
- data/lib/fselector/algo_discrete/ProbabilityRatio.rb +41 -0
- data/lib/fselector/algo_discrete/Random.rb +40 -0
- data/lib/fselector/algo_discrete/ReliefF_d.rb +173 -0
- data/lib/fselector/algo_discrete/Relief_d.rb +135 -0
- data/lib/fselector/algo_discrete/Sensitivity.rb +38 -0
- data/lib/fselector/algo_discrete/Specificity.rb +35 -0
- data/lib/fselector/base.rb +322 -0
- data/lib/fselector/base_continuous.rb +25 -0
- data/lib/fselector/base_discrete.rb +355 -0
- data/lib/fselector/ensemble.rb +181 -0
- data/lib/fselector/fileio.rb +455 -0
- data/lib/fselector/util.rb +707 -0
- metadata +86 -0
@@ -0,0 +1,40 @@
|
|
1
|
+
#
|
2
|
+
# FSelector: a Ruby gem for feature selection and ranking
|
3
|
+
#
|
4
|
+
module FSelector
|
5
|
+
#
|
6
|
+
# Random (Rand), no pratical use but can be used as a baseline
|
7
|
+
#
|
8
|
+
# Rand = rand numbers within [0..1)
|
9
|
+
#
|
10
|
+
# ref: [An extensive empirical study of feature selection metrics
|
11
|
+
# for text classification][url]
|
12
|
+
# [url]: http://dl.acm.org/citation.cfm?id=944974
|
13
|
+
#
|
14
|
+
class Random < BaseDiscrete
|
15
|
+
#
|
16
|
+
# initialize from an existing data structure
|
17
|
+
#
|
18
|
+
# @param [Integer] seed seed form random number
|
19
|
+
# generator. provided for reproducible results,
|
20
|
+
# otherwise use current time as seed
|
21
|
+
#
|
22
|
+
def initialize(seed=nil, data=nil)
|
23
|
+
super(data)
|
24
|
+
srand(seed) if seed
|
25
|
+
end
|
26
|
+
|
27
|
+
private
|
28
|
+
|
29
|
+
# calculate contribution of each feature (f) for each class (k)
|
30
|
+
def calc_contribution(f)
|
31
|
+
each_class do |k|
|
32
|
+
set_feature_score(f, k, rand)
|
33
|
+
end
|
34
|
+
end # calc_contribution
|
35
|
+
|
36
|
+
|
37
|
+
end # class
|
38
|
+
|
39
|
+
|
40
|
+
end # module
|
@@ -0,0 +1,173 @@
|
|
1
|
+
#
|
2
|
+
# FSelector: a Ruby gem for feature selection and ranking
|
3
|
+
#
|
4
|
+
module FSelector
|
5
|
+
# extended Relief algorithm for discrete feature (ReliefF_d)
|
6
|
+
#
|
7
|
+
# @note applicable to multi-class problem with missing data
|
8
|
+
#
|
9
|
+
# ref: [Estimating Attributes: Analysis and Extensions of RELIEF][url]
|
10
|
+
# [url]: http://www.springerlink.com/content/fp23jh2h0426ww45/
|
11
|
+
#
|
12
|
+
class ReliefF_d < BaseDiscrete
|
13
|
+
#
|
14
|
+
# new()
|
15
|
+
#
|
16
|
+
# @param [Integer] m number of samples to be used
|
17
|
+
# for estimating feature contribution. max can be
|
18
|
+
# the number of training samples
|
19
|
+
# @param [Integer] k number of k-nearest neighbor
|
20
|
+
# @param [Hash] data existing data structure
|
21
|
+
#
|
22
|
+
def initialize(m=nil, k=10, data=nil)
|
23
|
+
super(data)
|
24
|
+
@m = m # use all samples
|
25
|
+
@k = (k || 10) # default 10
|
26
|
+
end
|
27
|
+
|
28
|
+
private
|
29
|
+
|
30
|
+
# calculate contribution of each feature (f) across all classes
|
31
|
+
def calc_contribution(f)
|
32
|
+
score = 0.0
|
33
|
+
|
34
|
+
# use all samples if @m not provided
|
35
|
+
@m = get_sample_size if not @m
|
36
|
+
|
37
|
+
@m.times do
|
38
|
+
# pick a sample at random
|
39
|
+
rs, rk = pick_a_sample_at_random
|
40
|
+
|
41
|
+
# find k nearest neighbor for each class
|
42
|
+
nbrs = find_k_nearest_nb(rs, rk)
|
43
|
+
|
44
|
+
# calc contribution from neighbors
|
45
|
+
score += calc_score(f, rs, rk, nbrs)
|
46
|
+
end
|
47
|
+
|
48
|
+
s = score / @m
|
49
|
+
|
50
|
+
set_feature_score(f, :BEST, s)
|
51
|
+
end # calc_contribution
|
52
|
+
|
53
|
+
|
54
|
+
# pick a sample at random
|
55
|
+
def pick_a_sample_at_random
|
56
|
+
rk = get_classes[rand(get_classes.size)]
|
57
|
+
rks = get_data[rk]
|
58
|
+
|
59
|
+
[ rks[rand(rks.size)], rk ]
|
60
|
+
end # pick_a_sample_at_random
|
61
|
+
|
62
|
+
# # find k nearest neighbors of sample (rs) for each class
|
63
|
+
def find_k_nearest_nb(rs, rk)
|
64
|
+
nbrs = {}
|
65
|
+
|
66
|
+
each_class do |k|
|
67
|
+
res = []
|
68
|
+
|
69
|
+
get_data[k].each do |s|
|
70
|
+
next if s == rs # exclude self
|
71
|
+
|
72
|
+
d = diff_sample(rs, s, rk, k)
|
73
|
+
res << [d, s]
|
74
|
+
end
|
75
|
+
|
76
|
+
nbrs[k] = (res.sort { |x, y| x[0] <=> y[0] }[0...@k]).collect { |z| z[1] }
|
77
|
+
end
|
78
|
+
|
79
|
+
nbrs
|
80
|
+
end # find_k_nearest_nb
|
81
|
+
|
82
|
+
|
83
|
+
# difference between two samples
|
84
|
+
def diff_sample(s1, s2, k1, k2)
|
85
|
+
d = 0.0
|
86
|
+
|
87
|
+
each_feature do |f|
|
88
|
+
d += diff_feature(f, s1, s2, k1, k2)**2
|
89
|
+
end
|
90
|
+
|
91
|
+
d
|
92
|
+
end # diff_sample
|
93
|
+
|
94
|
+
|
95
|
+
# difference beween the feature (f) of two samples
|
96
|
+
def diff_feature(f, s1, s2, k1, k2)
|
97
|
+
d = 0.0
|
98
|
+
|
99
|
+
if s1.has_key?(f) and s2.has_key?(f) # no missing value
|
100
|
+
d = (s1[f] == s2[f]) ? 0.0 : 1.0
|
101
|
+
elsif not s1.has_key?(f) and not s2.has_key?(f) # two missing values
|
102
|
+
fvs = get_feature_values(f).uniq
|
103
|
+
fvs.each do |mv|
|
104
|
+
d -= calc_p(f, mv, k1)*calc_p(f, mv, k2)
|
105
|
+
end
|
106
|
+
d += 1
|
107
|
+
elsif not s1.has_key?(f) # s1: one missing value
|
108
|
+
# diff(f, s1, s2) = 1 - P(value(f, s2)|class(s1))
|
109
|
+
d = 1 - calc_p(f, s2[f], k1)
|
110
|
+
else # s2: one missing value
|
111
|
+
# diff(f, s1, s2) = 1 - P(value(f, s1)|class(s2))
|
112
|
+
d = 1 - calc_p(f, s1[f], k2)
|
113
|
+
end
|
114
|
+
|
115
|
+
d
|
116
|
+
end # diff_feature
|
117
|
+
|
118
|
+
|
119
|
+
# calc probability of missing value (mv)
|
120
|
+
def calc_p(f, mv, k)
|
121
|
+
# cache
|
122
|
+
if not @f2mvp
|
123
|
+
@f2mvp = {}
|
124
|
+
|
125
|
+
each_feature do |f|
|
126
|
+
@f2mvp[f] = {}
|
127
|
+
|
128
|
+
each_class do |k|
|
129
|
+
@f2mvp[f][k] = {}
|
130
|
+
|
131
|
+
fvs = get_feature_values(f).uniq
|
132
|
+
fvs.each do |v|
|
133
|
+
n = 0.0
|
134
|
+
|
135
|
+
get_data[k].each do |s|
|
136
|
+
n += 1 if s.has_key?(f) and s[f] == v
|
137
|
+
end
|
138
|
+
|
139
|
+
@f2mvp[f][k][v] = n/get_data[k].size
|
140
|
+
end
|
141
|
+
end
|
142
|
+
end
|
143
|
+
end
|
144
|
+
|
145
|
+
@f2mvp[f][k][mv]
|
146
|
+
end
|
147
|
+
|
148
|
+
|
149
|
+
# calc feature (f) contribution from neighbors
|
150
|
+
def calc_score(f, rs, rk, nbrs)
|
151
|
+
score = 0.0
|
152
|
+
|
153
|
+
nbrs.each do |k, nbs|
|
154
|
+
if k == rk # near hit
|
155
|
+
nbs.each do |s|
|
156
|
+
score -= (diff_feature(f, rs, s, rk, k)**2/nbs.size)
|
157
|
+
end
|
158
|
+
else # near_miss
|
159
|
+
nbs.each do |s|
|
160
|
+
score += (get_data[k].size/get_sample_size.to_f *
|
161
|
+
diff_feature(f, rs, s, rk, k)**2/nbs.size)
|
162
|
+
end
|
163
|
+
end
|
164
|
+
end
|
165
|
+
|
166
|
+
score
|
167
|
+
end
|
168
|
+
|
169
|
+
|
170
|
+
end # class
|
171
|
+
|
172
|
+
|
173
|
+
end # module
|
@@ -0,0 +1,135 @@
|
|
1
|
+
#
|
2
|
+
# FSelector: a Ruby gem for feature selection and ranking
|
3
|
+
#
|
4
|
+
module FSelector
|
5
|
+
#
|
6
|
+
# Relief algorithm for discrete feature (Relief_d)
|
7
|
+
#
|
8
|
+
# @note Relief applicable only to two-class problem without missing data
|
9
|
+
#
|
10
|
+
# ref: [The Feature Selection Problem: Traditional Methods
|
11
|
+
# and a New Algorithm][url]
|
12
|
+
# [url]: http://www.aaai.org/Papers/AAAI/1992/AAAI92-020.pdf
|
13
|
+
#
|
14
|
+
class Relief_d < BaseDiscrete
|
15
|
+
#
|
16
|
+
# new()
|
17
|
+
#
|
18
|
+
# @param [Integer] m number of samples to be used
|
19
|
+
# for estimating feature contribution. max can be
|
20
|
+
# the number of training samples
|
21
|
+
# @param [Hash] data existing data structure
|
22
|
+
#
|
23
|
+
def initialize(m=nil, data=nil)
|
24
|
+
super(data)
|
25
|
+
@m = m # default use all samples
|
26
|
+
end
|
27
|
+
|
28
|
+
private
|
29
|
+
|
30
|
+
# calculate contribution of each feature (f) across all classes
|
31
|
+
def calc_contribution(f)
|
32
|
+
if not get_classes.size == 2
|
33
|
+
abort "[#{__FILE__}@#{__LINE__}]: "+
|
34
|
+
"Relief applicable only to two-class problems without missing data"
|
35
|
+
end
|
36
|
+
|
37
|
+
# use all samples if @m not provided
|
38
|
+
@m = get_sample_size if not @m
|
39
|
+
|
40
|
+
k1, k2 = get_classes
|
41
|
+
score = 0.0
|
42
|
+
|
43
|
+
@m.times do
|
44
|
+
# pick a sample at random
|
45
|
+
rs, rk = pick_a_sample_at_random
|
46
|
+
|
47
|
+
# find the nearest neighbor for each class
|
48
|
+
nbrs = find_nearest_nb(rs, rk)
|
49
|
+
|
50
|
+
# calc contribution from neighbors
|
51
|
+
score += calc_score(f, rs, rk, nbrs)
|
52
|
+
end
|
53
|
+
|
54
|
+
s = score / @m
|
55
|
+
|
56
|
+
set_feature_score(f, :BEST, s)
|
57
|
+
end # calc_contribution
|
58
|
+
|
59
|
+
|
60
|
+
# pick a sample at random
|
61
|
+
def pick_a_sample_at_random
|
62
|
+
rk = get_classes[rand(get_classes.size)]
|
63
|
+
rks = get_data[rk]
|
64
|
+
|
65
|
+
[ rks[rand(rks.size)], rk ]
|
66
|
+
end # pick_a_sample_at_random
|
67
|
+
|
68
|
+
|
69
|
+
# find nearest neighbor sample for given sample (rs) within class (k)
|
70
|
+
def find_nearest_nb(rs, rk)
|
71
|
+
nbrs = {}
|
72
|
+
|
73
|
+
each_class do |k|
|
74
|
+
nb, dmin = nil, 999
|
75
|
+
get_data[k].each do |s|
|
76
|
+
next if s == rs # exclude self
|
77
|
+
d = diff_sample(rs, s)
|
78
|
+
if d < dmin
|
79
|
+
dmin = d
|
80
|
+
nb = s
|
81
|
+
end
|
82
|
+
end
|
83
|
+
|
84
|
+
nbrs[k] = nb
|
85
|
+
end
|
86
|
+
|
87
|
+
nbrs
|
88
|
+
end # find_nearest_nb
|
89
|
+
|
90
|
+
|
91
|
+
# difference between two samples
|
92
|
+
def diff_sample(s1, s2)
|
93
|
+
d = 0.0
|
94
|
+
|
95
|
+
each_feature do |f|
|
96
|
+
d += diff_feature(f, s1, s2)**2
|
97
|
+
end
|
98
|
+
|
99
|
+
d
|
100
|
+
end # diff_sample
|
101
|
+
|
102
|
+
|
103
|
+
# difference beween the feature (f) of two samples
|
104
|
+
def diff_feature(f, s1, s2)
|
105
|
+
d = 0.0
|
106
|
+
|
107
|
+
if not s1.has_key?(f) or not s2.has_key?(f)
|
108
|
+
abort "[#{__FILE__}@#{__LINE__}]: "+
|
109
|
+
"Relief does not allow missing values"
|
110
|
+
end
|
111
|
+
|
112
|
+
(s1[f] == s2[f]) ? 0.0 : 1.0
|
113
|
+
end # diff_feature
|
114
|
+
|
115
|
+
|
116
|
+
# calc feature (f) contribution from neighbors
|
117
|
+
def calc_score(f, rs, rk, nbrs)
|
118
|
+
score = 0.0
|
119
|
+
|
120
|
+
nbrs.each do |k, s|
|
121
|
+
if k == rk # near hit
|
122
|
+
score -= diff_feature(f, rs, s)**2
|
123
|
+
else # near_miss
|
124
|
+
score += diff_feature(f, rs, s)**2
|
125
|
+
end
|
126
|
+
end
|
127
|
+
|
128
|
+
score
|
129
|
+
end # calc_score
|
130
|
+
|
131
|
+
|
132
|
+
end # class
|
133
|
+
|
134
|
+
|
135
|
+
end # module
|
@@ -0,0 +1,38 @@
|
|
1
|
+
#
|
2
|
+
# FSelector: a Ruby gem for feature selection and ranking
|
3
|
+
#
|
4
|
+
module FSelector
|
5
|
+
#
|
6
|
+
# Sensitivity (SN)
|
7
|
+
#
|
8
|
+
# TP A
|
9
|
+
# SN = ------- = -----
|
10
|
+
# TP+FN A+C
|
11
|
+
#
|
12
|
+
class Sensitivity < BaseDiscrete
|
13
|
+
|
14
|
+
private
|
15
|
+
|
16
|
+
# calculate contribution of each feature (f) for each class (k)
|
17
|
+
def calc_contribution(f)
|
18
|
+
each_class do |k|
|
19
|
+
a, c = get_A(f, k), get_C(f, k)
|
20
|
+
|
21
|
+
s = a/(a+c)
|
22
|
+
|
23
|
+
set_feature_score(f, k, s)
|
24
|
+
end
|
25
|
+
end # calc_contribution
|
26
|
+
|
27
|
+
|
28
|
+
end # class
|
29
|
+
|
30
|
+
|
31
|
+
|
32
|
+
# shortcut so that you can use FSelector::SN instead of FSelector::Sensitivity
|
33
|
+
SN = Sensitivity
|
34
|
+
# Sensitivity, also known as Recall
|
35
|
+
Recall = Sensitivity
|
36
|
+
|
37
|
+
|
38
|
+
end # module
|
@@ -0,0 +1,35 @@
|
|
1
|
+
#
|
2
|
+
# FSelector: a Ruby gem for feature selection and ranking
|
3
|
+
#
|
4
|
+
module FSelector
|
5
|
+
#
|
6
|
+
# Specificity (SP)
|
7
|
+
#
|
8
|
+
# TN D
|
9
|
+
# SP = ------- = -----
|
10
|
+
# TN+FP B+D
|
11
|
+
#
|
12
|
+
class Specificity < BaseDiscrete
|
13
|
+
|
14
|
+
private
|
15
|
+
|
16
|
+
# calculate contribution of each feature (f) for each class (k)
|
17
|
+
def calc_contribution(f)
|
18
|
+
each_class do |k|
|
19
|
+
b, d = get_B(f, k), get_D(f, k)
|
20
|
+
|
21
|
+
s = d/(b+d)
|
22
|
+
|
23
|
+
set_feature_score(f, k, s)
|
24
|
+
end
|
25
|
+
end # calc_contribution
|
26
|
+
|
27
|
+
|
28
|
+
end # class
|
29
|
+
|
30
|
+
|
31
|
+
# shortcut so that you can use FSelector::SP instead of FSelector::Specificity
|
32
|
+
SP = Specificity
|
33
|
+
|
34
|
+
|
35
|
+
end # module
|
@@ -0,0 +1,322 @@
|
|
1
|
+
#
|
2
|
+
# FSelector: a Ruby gem for feature selection and ranking
|
3
|
+
#
|
4
|
+
module FSelector
|
5
|
+
#
|
6
|
+
# base ranking algorithm
|
7
|
+
#
|
8
|
+
class Base
|
9
|
+
# include FileIO
|
10
|
+
include FileIO
|
11
|
+
|
12
|
+
# initialize from an existing data structure
|
13
|
+
def initialize(data=nil)
|
14
|
+
@data = data
|
15
|
+
@opts = {} # store non-data information
|
16
|
+
end
|
17
|
+
|
18
|
+
|
19
|
+
#
|
20
|
+
# iterator for each class
|
21
|
+
#
|
22
|
+
# e.g.
|
23
|
+
# self.each_class do |k|
|
24
|
+
# puts k
|
25
|
+
# end
|
26
|
+
#
|
27
|
+
def each_class
|
28
|
+
if not block_given?
|
29
|
+
abort "[#{__FILE__}@#{__LINE__}]: "+
|
30
|
+
"block must be given!"
|
31
|
+
else
|
32
|
+
get_classes.each { |k| yield k }
|
33
|
+
end
|
34
|
+
end
|
35
|
+
|
36
|
+
|
37
|
+
#
|
38
|
+
# iterator for each feature
|
39
|
+
#
|
40
|
+
# e.g.
|
41
|
+
# self.each_feature do |f|
|
42
|
+
# puts f
|
43
|
+
# end
|
44
|
+
#
|
45
|
+
def each_feature
|
46
|
+
if not block_given?
|
47
|
+
abort "[#{__FILE__}@#{__LINE__}]: "+
|
48
|
+
"block must be given!"
|
49
|
+
else
|
50
|
+
get_features.each { |f| yield f }
|
51
|
+
end
|
52
|
+
end
|
53
|
+
|
54
|
+
|
55
|
+
#
|
56
|
+
# iterator for each sample with class label
|
57
|
+
#
|
58
|
+
# e.g.
|
59
|
+
# self.each_sample do |k, s|
|
60
|
+
# print k
|
61
|
+
# s.each { |f, v| ' '+v }
|
62
|
+
# puts
|
63
|
+
# end
|
64
|
+
#
|
65
|
+
def each_sample
|
66
|
+
if not block_given?
|
67
|
+
abort "[#{__FILE__}@#{__LINE__}]: "+
|
68
|
+
" block must be given!"
|
69
|
+
else
|
70
|
+
get_data.each do |k, samples|
|
71
|
+
samples.each { |s| yield k, s }
|
72
|
+
end
|
73
|
+
end
|
74
|
+
end
|
75
|
+
|
76
|
+
|
77
|
+
# get classes
|
78
|
+
def get_classes
|
79
|
+
@classes ||= @data.keys
|
80
|
+
end
|
81
|
+
|
82
|
+
|
83
|
+
# set classes
|
84
|
+
def set_classes(classes)
|
85
|
+
if classes and classes.class == Array
|
86
|
+
@classes = classes
|
87
|
+
else
|
88
|
+
abort "[#{__FILE__}@#{__LINE__}]: "+
|
89
|
+
"classes must be a Array object!"
|
90
|
+
end
|
91
|
+
end
|
92
|
+
|
93
|
+
|
94
|
+
# get unique features
|
95
|
+
def get_features
|
96
|
+
@features ||= @data.map { |x| x[1].map { |y| y.keys } }.flatten.uniq
|
97
|
+
end
|
98
|
+
|
99
|
+
|
100
|
+
#
|
101
|
+
# get feature values
|
102
|
+
#
|
103
|
+
# @param [Symbol] f feature of interest
|
104
|
+
#
|
105
|
+
def get_feature_values(f)
|
106
|
+
@fvs ||= {}
|
107
|
+
|
108
|
+
if not @fvs.has_key? f
|
109
|
+
@fvs[f] = []
|
110
|
+
each_sample do |k, s|
|
111
|
+
@fvs[f] << s[f] if s.has_key? f
|
112
|
+
end
|
113
|
+
end
|
114
|
+
|
115
|
+
@fvs[f]
|
116
|
+
end
|
117
|
+
|
118
|
+
# set features
|
119
|
+
def set_features(features)
|
120
|
+
if features and features.class == Array
|
121
|
+
@features = features
|
122
|
+
else
|
123
|
+
abort "[#{__FILE__}@#{__LINE__}]: "+
|
124
|
+
"features must be a Array object!"
|
125
|
+
end
|
126
|
+
end
|
127
|
+
|
128
|
+
|
129
|
+
# get data
|
130
|
+
def get_data
|
131
|
+
@data
|
132
|
+
end
|
133
|
+
|
134
|
+
# set data
|
135
|
+
def set_data(data)
|
136
|
+
if data and data.class == Hash
|
137
|
+
@data = data
|
138
|
+
# clear
|
139
|
+
@classes, @features, @fvs = nil, nil, nil
|
140
|
+
@scores, @ranks, @sz = nil, nil, nil
|
141
|
+
else
|
142
|
+
abort "[#{__FILE__}@#{__LINE__}]: "+
|
143
|
+
"data must be a Hash object!"
|
144
|
+
end
|
145
|
+
end
|
146
|
+
|
147
|
+
|
148
|
+
# get non-data information
|
149
|
+
def get_opt(key)
|
150
|
+
@opts.has_key?(key) ? @opts[key] : nil
|
151
|
+
end
|
152
|
+
|
153
|
+
|
154
|
+
# set non-data information as a key-value pair
|
155
|
+
def set_opt(key, value)
|
156
|
+
@opts[key] = value
|
157
|
+
end
|
158
|
+
|
159
|
+
|
160
|
+
# number of samples
|
161
|
+
def get_sample_size
|
162
|
+
@sz ||= get_data.values.flatten.size
|
163
|
+
end
|
164
|
+
|
165
|
+
|
166
|
+
#
|
167
|
+
# print feature scores
|
168
|
+
#
|
169
|
+
# @param [String] kclass class of interest
|
170
|
+
#
|
171
|
+
def print_feature_scores(feat=nil, kclass=nil)
|
172
|
+
scores = get_feature_scores
|
173
|
+
|
174
|
+
scores.each do |f, ks|
|
175
|
+
next if feat and feat != f
|
176
|
+
|
177
|
+
print "#{f} =>"
|
178
|
+
ks.each do |k, s|
|
179
|
+
if kclass
|
180
|
+
print " #{k}->#{s}" if k == kclass
|
181
|
+
else
|
182
|
+
print " #{k}->#{s}"
|
183
|
+
end
|
184
|
+
end
|
185
|
+
puts
|
186
|
+
end
|
187
|
+
end
|
188
|
+
|
189
|
+
|
190
|
+
# print feature ranks
|
191
|
+
def print_feature_ranks
|
192
|
+
ranks = get_feature_ranks
|
193
|
+
|
194
|
+
ranks.each do |f, r|
|
195
|
+
puts "#{f} => #{r}"
|
196
|
+
end
|
197
|
+
end
|
198
|
+
|
199
|
+
|
200
|
+
#
|
201
|
+
# get scores of all features for all classes
|
202
|
+
#
|
203
|
+
# @return [Hash] \{ feature =>
|
204
|
+
# \{ class_1 => score_1, class_2 => score_2, :BEST => score_best } }
|
205
|
+
#
|
206
|
+
def get_feature_scores
|
207
|
+
return @scores if @scores # already done
|
208
|
+
|
209
|
+
each_feature do |f|
|
210
|
+
calc_contribution(f)
|
211
|
+
end
|
212
|
+
|
213
|
+
# best score for feature
|
214
|
+
@scores.each do |f, ks|
|
215
|
+
# the larger, the better
|
216
|
+
@scores[f][:BEST] = ks.values.max
|
217
|
+
end
|
218
|
+
#@scores.each { |x,v| puts "#{x} => #{v[:BEST]}" }
|
219
|
+
|
220
|
+
@scores
|
221
|
+
end
|
222
|
+
|
223
|
+
|
224
|
+
# set feature (f) score (f) for class (k)
|
225
|
+
def set_feature_score(f, k, s)
|
226
|
+
@scores ||= {}
|
227
|
+
@scores[f] ||= {}
|
228
|
+
@scores[f][k] = s
|
229
|
+
end
|
230
|
+
|
231
|
+
#
|
232
|
+
# get the ranked features based on their best scores
|
233
|
+
#
|
234
|
+
# @return [Hash] feature ranks
|
235
|
+
#
|
236
|
+
def get_feature_ranks
|
237
|
+
return @ranks if @ranks # already done
|
238
|
+
|
239
|
+
scores = get_feature_scores
|
240
|
+
|
241
|
+
# get the ranked features
|
242
|
+
@ranks = {} # feature => rank
|
243
|
+
|
244
|
+
# the larger, the better
|
245
|
+
sorted_features = scores.keys.sort do |x,y|
|
246
|
+
scores[y][:BEST] <=> scores[x][:BEST]
|
247
|
+
end
|
248
|
+
|
249
|
+
sorted_features.each_with_index do |sf, si|
|
250
|
+
@ranks[sf] = si+1
|
251
|
+
end
|
252
|
+
|
253
|
+
@ranks
|
254
|
+
end
|
255
|
+
|
256
|
+
|
257
|
+
#
|
258
|
+
# reconstruct data with feature scores satisfying cutoff
|
259
|
+
#
|
260
|
+
# @param [String] criterion
|
261
|
+
# valid criterion can be '>0.5', '>= 0.4', '==2', '<=1' or '<0.2'
|
262
|
+
# @param [Hash] my_scores
|
263
|
+
# user customized feature scores
|
264
|
+
# @return [Hash] data after feature selection
|
265
|
+
# @note data structure will be altered
|
266
|
+
#
|
267
|
+
def select_data_by_score!(criterion, my_scores=nil)
|
268
|
+
# user scores or internal scores
|
269
|
+
scores = my_scores || get_feature_scores
|
270
|
+
|
271
|
+
my_data = {}
|
272
|
+
|
273
|
+
each_sample do |k, s|
|
274
|
+
my_data[k] ||= []
|
275
|
+
my_s = {}
|
276
|
+
|
277
|
+
s.each do |f, v|
|
278
|
+
my_s[f] = v if eval("#{scores[f][:BEST]} #{criterion}")
|
279
|
+
end
|
280
|
+
|
281
|
+
my_data[k] << my_s if not my_s.empty?
|
282
|
+
end
|
283
|
+
|
284
|
+
set_data(my_data)
|
285
|
+
end
|
286
|
+
|
287
|
+
|
288
|
+
#
|
289
|
+
# reconstruct data by rank
|
290
|
+
#
|
291
|
+
# @param [String] criterion
|
292
|
+
# valid criterion can be '>11', '>= 10', '==1', '<=10' or '<20'
|
293
|
+
# @param [Hash] my_ranks
|
294
|
+
# user customized feature ranks
|
295
|
+
# @return [Hash] data after feature selection
|
296
|
+
# @note data structure will be altered
|
297
|
+
#
|
298
|
+
def select_data_by_rank!(criterion, my_ranks=nil)
|
299
|
+
# user ranks or internal ranks
|
300
|
+
ranks = my_ranks || get_feature_ranks
|
301
|
+
|
302
|
+
my_data = {}
|
303
|
+
|
304
|
+
each_sample do |k, s|
|
305
|
+
my_data[k] ||= []
|
306
|
+
my_s = {}
|
307
|
+
|
308
|
+
s.each do |f,v|
|
309
|
+
my_s[f] = v if eval("#{ranks[f]} #{criterion}")
|
310
|
+
end
|
311
|
+
|
312
|
+
my_data[k] << my_s if not my_s.empty?
|
313
|
+
end
|
314
|
+
|
315
|
+
set_data(my_data)
|
316
|
+
end
|
317
|
+
|
318
|
+
|
319
|
+
end # class
|
320
|
+
|
321
|
+
|
322
|
+
end # module
|