数据集Node.js中的最近邻居

我有一个将数据保存到MongoDB的Node.js应用程序。 给定一个文件，我想在数据库中find最相似的文件。

我的想法是实现某种最近邻algorithm，将所有logging作为训练序列，并返回最相似的文档（包括这两个文档相似程度的某种百分比）。

例如在我的数据库中有这些logging…

{ name: "Bill", age: 10, pc: "Mac", ip: "68.23.13.8" } { name: "Alice", age: 22, pc: "Windows", ip: "193.186.11.3" } { name: "Bob", age: 12, pc: "Windows", ip: "56.89.22.1" } 

…我想find最接近这个文件的文件

 { name: "Tom", age: 10, pc: "Mac", ip: "68.23.13.10" } // algorithm returns "Bill", .76 

是否有任何节点模块/实现采取任何forms的对象/参数，并返回最近的邻居？

Answer 1

这里是一些示例代码。 它假定您可以在每个请求上运行search。 如果要修改它，请确保所有相似性函数返回0到1之间的数字。

 function tokenize(string) { var tokens = []; for (var i = 0; i < string.length-1; i++) { tokens.push(string.substr(i,2)); } return tokens.sort(); } function intersect(a, b) { var ai=0, bi=0; var result = new Array(); while( ai < a.length && bi < b.length ) { if (a[ai] < b[bi] ){ ai++; } else if (a[ai] > b[bi] ){ bi++; } else /* they're equal */ { result.push(a[ai]); ai++; bi++; } } return result; } function sum(items) { var sum = 0; for (var i = 0; i < items.length; i++) { sum += items[i]; } return sum; } function wordSimilarity(a, b) { var left = tokenize(a); var right = tokenize(b); var middle = intersect(left, right); return (2*middle.length) / (left.length + right.length); } function ipSimilarity(a, b) { var left = a.split('.'); var right = b.split('.'); var diffs = []; for (var i = 0; i < 4; i++) { var diff1 = 255-left[i]; var diff2 = 255-right[i]; var diff = Math.abs(diff2-diff1); diffs[i] = diff; } var distance = sum(diffs)/(255*4); return 1 - distance; } function ageSimilarity(a, b) { var maxAge = 100; var diff1 = maxAge-a; var diff2 = maxAge-b; var diff = Math.abs(diff2-diff1); var distance = diff / maxAge; return 1-distance; } function recordSimilarity(a, b) { var fields = [ {name:'name', measure:wordSimilarity}, {name:'age', measure:ageSimilarity}, {name:'pc', measure:wordSimilarity}, {name:'ip', measure:ipSimilarity} ]; var sum = 0; for (var i = 0; i < fields.length; i++) { var field = fields[i]; var name = field.name; var measure = field.measure; var sim = measure(a[name], b[name]); sum += sim; } return sum / fields.length; } function findMostSimilar(items, query) { var maxSim = 0; var result = null; for (var i = 0; i < items.length; i++) { var item = items[i]; var sim = recordSimilarity(item, query); if (sim > maxSim) { maxSim = sim; result = item; } } return result } var items = [ { name: "Bill", age: 10, pc: "Mac", ip: "68.23.13.8" }, { name: "Alice", age: 22, pc: "Windows", ip: "193.186.11.3" }, { name: "Bob", age: 12, pc: "Windows", ip: "56.89.22.1" } ]; var query = { name: "Tom", age: 10, pc: "Mac", ip: "68.23.13.10" }; var result = findMostSimilar(items, query); console.log(result); 

Answer 2

一个简单的方法是计算两个文档之间的差异，差异越大，距离越大。 你可以使用最大可能的差异来标准化差异，这应该给你相对的距离，你可以相互比较。

看看这个问题来计算JSON文档上的差异。

JSON对象的增量编码