/
diana.js
122 lines (113 loc) · 2.3 KB
/
diana.js
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
/**
* DIvisive ANAlysis Clustering
*/
export default class DIANA {
// https://www.slideshare.net/sesejun/datamining-8th-hclustering
constructor() {}
/**
* Number of clusters
*
* @type {number}
*/
get size() {
return this._tree.leafs.length
}
/**
* Initialize model.
*
* @param {Array<Array<number>>} datas Training data
*/
init(datas) {
this._x = datas
this._tree = {
idx: datas.map((_, i) => i),
children: [],
get leafs() {
return this.children.length === 0 ? [this] : this.children.reduce((c, v) => c.concat(v.leafs), [])
},
}
}
_distance(a, b) {
return Math.sqrt(a.reduce((s, v, i) => s + (v - b[i]) ** 2, 0))
}
_v(i, v, s) {
let a = 0
for (let k = 0; k < v.length; k++) {
if (v[k] !== i && !s.includes(v[k])) {
a += this._distance(this._x[i], this._x[v[k]])
}
}
a = a / (v.length - s.length - 1)
if (s.length > 0) {
let b = 0
for (let k = 0; k < s.length; k++) {
b += this._distance(this._x[i], this._x[s[k]])
}
a -= b / s.length
}
return a
}
/**
* Fit model.
*/
fit() {
for (const leaf of this._tree.leafs) {
if (leaf.idx.length === 1) {
continue
}
const s = []
while (s.length < leaf.idx.length) {
let max_v = -Infinity
let max_i = -1
for (let i = 0; i < leaf.idx.length; i++) {
if (s.includes(leaf.idx[i])) continue
const a = this._v(leaf.idx[i], leaf.idx, s)
if (max_v < a) {
max_v = a
max_i = leaf.idx[i]
}
}
if (max_v <= 0) break
s.push(max_i)
}
if (0 < s.length && s.length < leaf.idx.length) {
const s0 = leaf.idx.filter(a => !s.includes(a))
leaf.children.push(
{
idx: s,
children: [],
get leafs() {
return this.children.length === 0
? [this]
: this.children.reduce((c, v) => c.concat(v.leafs), [])
},
},
{
idx: s0,
children: [],
get leafs() {
return this.children.length === 0
? [this]
: this.children.reduce((c, v) => c.concat(v.leafs), [])
},
}
)
}
}
}
/**
* Returns predicted categories.
*
* @returns {number[]} Predicted values
*/
predict() {
const p = []
const leafs = this._tree.leafs
for (let k = 0; k < leafs.length; k++) {
for (const i of leafs[k].idx) {
p[i] = k
}
}
return p
}
}