@antv/g6
Version:
graph visualization frame work
71 lines (66 loc) • 1.85 kB
JavaScript
const Numeric = require('numericjs');
class MDS {
// getDefaultCfgs() {
// return {
// distances: null, // 停止迭代的最大迭代数
// demension: 2 // 中心点,默认为数据中第一个点
// };
// }
constructor(params) {
/**
* distance matrix
* @type {array}
*/
this.distances = params.distances;
/**
* dimensions
* @type {number}
*/
this.dimension = params.dimension || 2;
/**
* link distance
* @type {number}
*/
this.linkDistance = params.linkDistance;
}
layout() {
const self = this;
const dimension = self.dimension;
const distances = self.distances;
const linkDistance = self.linkDistance;
// square distances
const M = Numeric.mul(-0.5, Numeric.pow(distances, 2));
// double centre the rows/columns
function mean(A) { return Numeric.div(Numeric.add.apply(null, A), A.length); }
const rowMeans = mean(M),
colMeans = mean(Numeric.transpose(M)),
totalMean = mean(rowMeans);
for (let i = 0; i < M.length; ++i) {
for (let j = 0; j < M[0].length; ++j) {
M[i][j] += totalMean - rowMeans[i] - colMeans[j];
}
}
// take the SVD of the double centred matrix, and return the
// points from it
let ret;
let res = [];
try {
ret = Numeric.svd(M);
} catch (e) {
const length = distances.length;
for (let i = 0; i < length; i++) {
const x = Math.random() * linkDistance;
const y = Math.random() * linkDistance;
res.push([ x, y ]);
}
}
if (res.length === 0) {
const eigenValues = Numeric.sqrt(ret.S);
res = ret.U.map(function(row) {
return Numeric.mul(row, eigenValues).splice(0, dimension);
});
}
return res;
}
}
module.exports = MDS;