使用Cytoscape.js在所有节点上执行集中功能 [英] Perform Centrality Functions on all Nodes using Cytoscape.js

问题描述

我需要计算图上每个节点的程度，紧密度和中间度。我目前正在 cy.ready（）事件之后的每个节点上使用Cytoscape.js内置的函数。但是，由于图形非常大（250个以上的节点，650个以上的连接），因此计算时间太长。有人可以建议一种更有效的方法吗？

I need to calculate degree, closeness and betweenness centrality for every node on a graph. I'm currently using the functions built into Cytoscape.js on each node after the cy.ready() event. However, as the graphs are quite large (250+ Nodes, 650+ Connections) it's taking too long to compute. Can anyone suggest a more efficient method?

var calculateSNA = function() {
  // Don't run if already set...
  if(data.sna) return false
  console.log('Running SNA')

  _.map(nodes, function(node) {
    var target = cy.nodes('#' + node.data.id)
    node.data.sna  = {
      degreeCentrality: cy.$().dc({ root: target }).degree,
      closenessCentrality: cy.$().cc({ root: target }),
      betweennessCentrality: cy.$().bc().betweenness(target)
    }
    return node
  })

  // Add SNA data to local storage
  Data.add({
    sna: true
  })
  Node.set(nodes)

  console.log('SNA complete')
}


cy.ready(function(event) {
  console.log('cy.ready()')
  calculateSNA()
})

推荐答案

运行一次算法，而不是运行N次。然后只需查询结果即可：

Run the algorithm once, rather than running it N times. Then just query the result:

let ccn = cy.elements().closenessCentralityNormalized({ /* my options */ });

cy.nodes().forEach( n => {
  n.data({
    ccn: ccn.closeness( n )
  });
} );

使用中心化算法的规范化版本，除非您有充分的理由这样做。例如，如果您将一个图形中的结果与另一图形中的结果进行比较，则只有规范化版本才真正有意义。

Use normalised versions of centrality algorithms unless you have a good reason to do otherwise. Only the normalised versions really mean anything if you compare results in one graph to results in another graph, for example.

这篇关于使用Cytoscape.js在所有节点上执行集中功能的文章就介绍到这了，希望我们推荐的答案对大家有所帮助，也希望大家多多支持IT屋！