forked from ryanbressler/CloudForest
-
Notifications
You must be signed in to change notification settings - Fork 0
/
libsvm.go
198 lines (160 loc) · 4.13 KB
/
libsvm.go
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
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
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
package CloudForest
import (
"bufio"
"encoding/csv"
"fmt"
"io"
"log"
"strconv"
"strings"
)
func ParseLibSVM(input io.Reader) *FeatureMatrix {
reader := bufio.NewReader(input)
data := make([]Feature, 0, 100)
lookup := make(map[string]int, 0)
labels := make([]string, 0, 0)
i := 0
ncases := 0
for {
ncases++
line, err := reader.ReadString('\n')
if err == io.EOF {
break
} else if err != nil {
log.Print("Error:", err)
return nil
}
vals := strings.Fields(line)
if i == 0 {
name := "0"
lookup[name] = 0
if strings.Contains(vals[0], ".") {
//looks like a float...add dense float64 feature regression
data = append(data, &DenseNumFeature{
make([]float64, 0, 0),
make([]bool, 0, 0),
name,
false})
} else {
//doesn't look like a float...add dense catagorical
data = append(data, &DenseCatFeature{
&CatMap{make(map[string]int, 0),
make([]string, 0, 0)},
make([]int, 0, 0),
make([]bool, 0, 0),
name,
false,
false})
}
}
data[0].Append(vals[0])
//pad existing features
for _, f := range data[1:] {
f.Append("0")
}
for _, v := range vals[1:] {
parts := strings.Split(v, ":")
xi, err := strconv.Atoi(parts[0])
if err != nil {
log.Print("Atoi error: ", err, " Line ", i, " Parsing: ", v)
}
//pad out the data to include this feature
for xi >= len(data) {
name := fmt.Sprintf("%v", len(data))
lookup[name] = len(data)
data = append(data, &DenseNumFeature{
make([]float64, ncases, ncases),
make([]bool, ncases, ncases),
name,
false})
}
data[xi].PutStr(i, parts[1])
}
label := fmt.Sprintf("%v", i)
labels = append(labels, label)
i++
}
fm := &FeatureMatrix{data, lookup, labels}
return fm
}
func WriteLibSvm(data *FeatureMatrix, targetn string, outfile io.Writer) error {
targeti, ok := data.Map[targetn]
if !ok {
return fmt.Errorf("Target '%v' not found in data.", targetn)
}
target := data.Data[targeti]
//data.Data = append(data.Data[:targeti], data.Data[targeti+1:]...)
noTargetFm := &FeatureMatrix{make([]Feature, 0, len(data.Data)), make(map[string]int), data.CaseLabels}
for i, f := range data.Data {
if i != targeti {
noTargetFm.Map[f.GetName()] = len(noTargetFm.Data)
noTargetFm.Data = append(noTargetFm.Data, f.Copy())
}
}
noTargetFm.ImputeMissing()
encodedfm := noTargetFm.EncodeToNum()
oucsv := csv.NewWriter(outfile)
oucsv.Comma = ' '
for i := 0; i < target.Length(); i++ {
entries := make([]string, 0, 10)
switch target.(type) {
case NumFeature:
entries = append(entries, target.GetStr(i))
case CatFeature:
entries = append(entries, fmt.Sprintf("%v", target.(CatFeature).Geti(i)))
}
for j, f := range encodedfm.Data {
v := f.(NumFeature).Get(i)
if v != 0.0 {
entries = append(entries, fmt.Sprintf("%v:%v", j+1, v))
}
}
//fmt.Println(entries)
err := oucsv.Write(entries)
if err != nil {
return err
}
}
oucsv.Flush()
return nil
}
func WriteLibSvmCases(data *FeatureMatrix, cases []int, targetn string, outfile io.Writer) error {
targeti, ok := data.Map[targetn]
if !ok {
return fmt.Errorf("Target '%v' not found in data.", targetn)
}
target := data.Data[targeti]
noTargetFm := &FeatureMatrix{make([]Feature, 0, len(data.Data)), make(map[string]int), data.CaseLabels}
for i, f := range data.Data {
if i != targeti {
noTargetFm.Map[f.GetName()] = len(noTargetFm.Data)
noTargetFm.Data = append(noTargetFm.Data, f.Copy())
}
}
noTargetFm.ImputeMissing()
encodedfm := noTargetFm.EncodeToNum()
oucsv := csv.NewWriter(outfile)
oucsv.Comma = ' '
for _, i := range cases {
entries := make([]string, 0, 10)
switch target.(type) {
case NumFeature:
entries = append(entries, fmt.Sprintf("%g", target.(NumFeature).Get(i)))
case CatFeature:
entries = append(entries, fmt.Sprintf("%v", target.(CatFeature).Geti(i)))
}
for j, f := range encodedfm.Data {
v := f.(NumFeature).Get(i)
if v != 0.0 {
entries = append(entries, fmt.Sprintf("%v:%v", j+1, v))
}
}
//fmt.Println(entries)
err := oucsv.Write(entries)
if err != nil {
return err
}
}
oucsv.Flush()
return nil
}