【问题标题】:Compare Strings Javascript Return %of Likely比较字符串 Javascript 返回 %of Likely
【发布时间】:2012-05-15 11:15:07
【问题描述】:

我正在寻找一个 JavaScript 函数,它可以比较两个字符串并返回它们相似的可能性。我看过 soundex 但这对于多词字符串或非名称来说并不是很好。我正在寻找这样的功能:

    function compare(strA,strB){
    
    }
    
    compare("Apples","apple") = Some X Percentage.

该函数适用于所有类型的字符串,包括数字、多字值和名称。也许我可以使用一个简单的算法?

最终这些都没有达到我的目的,所以我使用了这个:

     function compare(c, u) {
            var incept = false;
            var ca = c.split(",");
            u = clean(u);
            //ca = correct answer array (Collection of all correct answer)
            //caa = a single correct answer word array (collection of words of a single correct answer)
            //u = array of user answer words cleaned using custom clean function
            for (var z = 0; z < ca.length; z++) {
                caa = $.trim(ca[z]).split(" ");
                var pc = 0;
                for (var x = 0; x < caa.length; x++) {
                    for (var y = 0; y < u.length; y++) {
                        if (soundex(u[y]) != null && soundex(caa[x]) != null) {
                            if (soundex(u[y]) == soundex(caa[x])) {
                                pc = pc + 1;
                            }
                        }
                        else {
                            if (u[y].indexOf(caa[x]) > -1) {
                                pc = pc + 1;
                            }
                        }
                    }
                }
                if ((pc / caa.length) > 0.5) {
                    return true;
                }
            }
            return false;
        }
        
        // create object listing the SOUNDEX values for each letter
        // -1 indicates that the letter is not coded, but is used for coding
        //  0 indicates that the letter is omitted for modern census archives
        //                              but acts like -1 for older census archives
        //  1 is for BFPV
        //  2 is for CGJKQSXZ
        //  3 is for DT
        //  4 is for L
        //  5 is for MN my home state
        //  6 is for R
        function makesoundex() {
            this.a = -1
            this.b = 1
            this.c = 2
            this.d = 3
            this.e = -1
            this.f = 1
            this.g = 2
            this.h = 0
            this.i = -1
            this.j = 2
            this.k = 2
            this.l = 4
            this.m = 5
            this.n = 5
            this.o = -1
            this.p = 1
            this.q = 2
            this.r = 6
            this.s = 2
            this.t = 3
            this.u = -1
            this.v = 1
            this.w = 0
            this.x = 2
            this.y = -1
            this.z = 2
        }
        
        var sndx = new makesoundex()
        
        // check to see that the input is valid
        function isSurname(name) {
            if (name == "" || name == null) {
                return false
            } else {
                for (var i = 0; i < name.length; i++) {
                    var letter = name.charAt(i)
                    if (!(letter >= 'a' && letter <= 'z' || letter >= 'A' && letter <= 'Z')) {
                        return false
                    }
                }
            }
            return true
        }
        
        // Collapse out directly adjacent sounds
        // 1. Assume that surname.length>=1
        // 2. Assume that surname contains only lowercase letters
        function collapse(surname) {
            if (surname.length == 1) {
                return surname
            }
            var right = collapse(surname.substring(1, surname.length))
            if (sndx[surname.charAt(0)] == sndx[right.charAt(0)]) {
                return surname.charAt(0) + right.substring(1, right.length)
            }
            return surname.charAt(0) + right
        }
        
        // Collapse out directly adjacent sounds using the new National Archives method
        // 1. Assume that surname.length>=1
        // 2. Assume that surname contains only lowercase letters
        // 3. H and W are completely ignored
        function omit(surname) {
            if (surname.length == 1) {
                return surname
            }
            var right = omit(surname.substring(1, surname.length))
            if (!sndx[right.charAt(0)]) {
                return surname.charAt(0) + right.substring(1, right.length)
            }
            return surname.charAt(0) + right
        }
        
        // Output the coded sequence
        function output_sequence(seq) {
            var output = seq.charAt(0).toUpperCase() // Retain first letter
            output += "-" // Separate letter with a dash
            var stage2 = seq.substring(1, seq.length)
            var count = 0
            for (var i = 0; i < stage2.length && count < 3; i++) {
                if (sndx[stage2.charAt(i)] > 0) {
                    output += sndx[stage2.charAt(i)]
                    count++
                }
            }
            for (; count < 3; count++) {
                output += "0"
            }
            return output
        }
        
        // Compute the SOUNDEX code for the surname
        function soundex(value) {
            if (!isSurname(value)) {
                return null
            }
            var stage1 = collapse(value.toLowerCase())
            //form.result.value=output_sequence(stage1);
        
            var stage1 = omit(value.toLowerCase())
            var stage2 = collapse(stage1)
            return output_sequence(stage2);
        
        }
        
        function clean(u) {
            var u = u.replace(/\,/g, "");
            u = u.toLowerCase().split(" ");
            var cw = ["ARRAY OF WORDS TO BE EXCLUDED FROM COMPARISON"];
            var n = [];
            for (var y = 0; y < u.length; y++) {
                var test = false;
                for (var z = 0; z < cw.length; z++) {
                    if (u[y] != "" && u[y] != cw[z]) {
                        test = true;
                        break;
                    }
                }
                if (test) {
        //Don't use & or $ in comparison
                    var val = u[y].replace("$", "").replace("&", "");
                    n.push(val);
                }
            }
            return n;
        }

【问题讨论】:

标签: javascript string algorithm comparison


【解决方案1】:

这是一个基于 Levenshtein 距离 https://en.wikipedia.org/wiki/Levenshtein_distance 的答案

function similarity(s1, s2) {
  var longer = s1;
  var shorter = s2;
  if (s1.length < s2.length) {
    longer = s2;
    shorter = s1;
  }
  var longerLength = longer.length;
  if (longerLength == 0) {
    return 1.0;
  }
  return (longerLength - editDistance(longer, shorter)) / parseFloat(longerLength);
}

用于计算编辑距离

function editDistance(s1, s2) {
  s1 = s1.toLowerCase();
  s2 = s2.toLowerCase();

  var costs = new Array();
  for (var i = 0; i <= s1.length; i++) {
    var lastValue = i;
    for (var j = 0; j <= s2.length; j++) {
      if (i == 0)
        costs[j] = j;
      else {
        if (j > 0) {
          var newValue = costs[j - 1];
          if (s1.charAt(i - 1) != s2.charAt(j - 1))
            newValue = Math.min(Math.min(newValue, lastValue),
              costs[j]) + 1;
          costs[j - 1] = lastValue;
          lastValue = newValue;
        }
      }
    }
    if (i > 0)
      costs[s2.length] = lastValue;
  }
  return costs[s2.length];
}

用法

similarity('Stack Overflow','Stack Ovrflw')

返回 0.8571428571428571


你可以在下面玩它:

function checkSimilarity(){
  var str1 = document.getElementById("lhsInput").value;
  var str2 = document.getElementById("rhsInput").value;
  document.getElementById("output").innerHTML = similarity(str1, str2);
}

function similarity(s1, s2) {
      var longer = s1;
      var shorter = s2;
      if (s1.length < s2.length) {
        longer = s2;
        shorter = s1;
      }
      var longerLength = longer.length;
      if (longerLength == 0) {
        return 1.0;
      }
      return (longerLength - editDistance(longer, shorter)) / parseFloat(longerLength);
    }

    function editDistance(s1, s2) {
      s1 = s1.toLowerCase();
      s2 = s2.toLowerCase();

      var costs = new Array();
      for (var i = 0; i <= s1.length; i++) {
        var lastValue = i;
        for (var j = 0; j <= s2.length; j++) {
          if (i == 0)
            costs[j] = j;
          else {
            if (j > 0) {
              var newValue = costs[j - 1];
              if (s1.charAt(i - 1) != s2.charAt(j - 1))
                newValue = Math.min(Math.min(newValue, lastValue),
                  costs[j]) + 1;
              costs[j - 1] = lastValue;
              lastValue = newValue;
            }
          }
        }
        if (i > 0)
          costs[s2.length] = lastValue;
      }
      return costs[s2.length];
    }
<div><label for="lhsInput">String 1:</label> <input type="text" id="lhsInput" oninput="checkSimilarity()" /></div>
<div><label for="rhsInput">String 2:</label> <input type="text" id="rhsInput" oninput="checkSimilarity()" /></div>
<div>Match: <span id="output">No Input</span></div>

【讨论】:

  • 对几个词的改进:var similarity2 = function(s1, s2){ var split1 = s1.split(' '); var split2 = s2.split(' ');变量总和 = 0;变量最大值 = 0;变量温度 = 0; for(var i=0; i
  • @ vervord1234是上面的方法适用于字符串吗
  • 它确实适用于没有附加语义的字符串。请尝试运行内联代码 sn-p(感谢 David)。当我输入上述字符串时,我得到了 0.17857142857142858 的相似度。
  • @hyperfkcb 他正在实施编辑距离算法,该算法计算错误位置的字符数(或多或少),因此为了计算百分比,他采用了更长的编辑距离值(longerLength)并做 (longerLength - editDistance) / longLength.
  • 但是对于长字符串来说太慢了。
【解决方案2】:

使用 this 库来进行字符串相似性对我来说就像是一种魅力!

这是示例 -

var similarity = stringSimilarity.compareTwoStrings("Apples","apple");    // => 0.88

【讨论】:

  • 这很好,除了 stringSimilarity 有一个名为 lodash 的依赖项,它包含超过 1,000 个文件被放入您的项目中,这样您就可以获得字符串相似度。
  • 是的,在本地添加包时会发生这种情况。但是,我们可以使用 CDN 来获得更小的包大小。这是 CDN 链接 - jsdelivr.com/package/npm/lodash - jsdelivr.com/package/npm/string-similarity
  • 他们已经移除了大部分依赖,包括 lodash
【解决方案3】:

这是一个非常简单的函数,它进行比较并根据等价返回一个百分比。虽然它尚未针对所有可能的情况进行测试,但它可能会帮助您入门。

function similar(a,b) {
    var equivalency = 0;
    var minLength = (a.length > b.length) ? b.length : a.length;    
    var maxLength = (a.length < b.length) ? b.length : a.length;    
    for(var i = 0; i < minLength; i++) {
        if(a[i] == b[i]) {
            equivalency++;
        }
    }
    

    var weight = equivalency / maxLength;
    return (weight * 100) + "%";
}
alert(similar("test","tes"));   // 75%
alert(similar("test","test"));  // 100%
alert(similar("test","testt")); // 80%
alert(similar("test","tess"));  // 75%

【讨论】:

  • 这里的问题是 "test" 和 "test" 返回 0%,我们知道这不是真的。
【解决方案4】:

查找两个字符串之间的相似度;我们可以使用不止一种或两种方法,但我更倾向于使用 'Dice's Coefficient' 。哪个更好!据我所知,比使用 'Levenshtein distance'

使用来自 npm 的这个 'string-similarity' 包,您将能够处理我上面所说的内容。

一些简单的用法示例是

var stringSimilarity = require('string-similarity');

var similarity = stringSimilarity.compareTwoStrings('healed', 'sealed'); 

var matches = stringSimilarity.findBestMatch('healed', ['edward', 'sealed', 'theatre']);

更多信息请访问上面给出的链接。谢谢。

【讨论】:

【解决方案5】:

我很快就写了一篇,可能足以满足您的目的:

function Compare(strA,strB){
    for(var result = 0, i = strA.length; i--;){
        if(typeof strB[i] == 'undefined' || strA[i] == strB[i]);
        else if(strA[i].toLowerCase() == strB[i].toLowerCase())
            result++;
        else
            result += 4;
    }
    return 1 - (result + 4*Math.abs(strA.length - strB.length))/(2*(strA.length+strB.length));
}

这对相同但大小写不同的字符 1 的权重与完全不同或缺失的字符的权重相同。它返回一个介于 0 和 1 之间的数字,1 表示字符串相同。 0 表示它们没有相似之处。例子:

Compare("Apple", "Apple")    // 1
Compare("Apples", "Apple")   // 0.8181818181818181
Compare("Apples", "apple")   // 0.7727272727272727
Compare("a", "A")            // 0.75
Compare("Apples", "appppp")  // 0.45833333333333337
Compare("a", "b")            // 0

【讨论】:

  • 不太准确:Compare("Apple", "zApple") = 0.07 ,而 Compare("Apple", "Applez") = 0.84
  • @Kousha ,它的位置。 “Apple”和“zApple”只有一个字母相同(第二个p)。
  • @Paulpro Apple 和 zApple 在逻辑上共有五个字母。这是您的实施错误。 Apple、zApple、Applez 类似。
  • @Kousha,根据这个算法,zApple 并不相似,因为它是有位置的。这不会使算法不正确。
  • @Paulpro:这不会让你的算法不正确,但会导致这个问题的答案很差......
【解决方案6】:

来自PHP.js library的函数similar_text怎么样?

它基于一个带有the same name的PHP函数。

function similar_text (first, second) {
    // Calculates the similarity between two strings  
    // discuss at: http://phpjs.org/functions/similar_text

    if (first === null || second === null || typeof first === 'undefined' || typeof second === 'undefined') {
        return 0;
    }

    first += '';
    second += '';

    var pos1 = 0,
        pos2 = 0,
        max = 0,
        firstLength = first.length,
        secondLength = second.length,
        p, q, l, sum;

    max = 0;

    for (p = 0; p < firstLength; p++) {
        for (q = 0; q < secondLength; q++) {
            for (l = 0;
            (p + l < firstLength) && (q + l < secondLength) && (first.charAt(p + l) === second.charAt(q + l)); l++);
            if (l > max) {
                max = l;
                pos1 = p;
                pos2 = q;
            }
        }
    }

    sum = max;

    if (sum) {
        if (pos1 && pos2) {
            sum += this.similar_text(first.substr(0, pos2), second.substr(0, pos2));
        }

        if ((pos1 + max < firstLength) && (pos2 + max < secondLength)) {
            sum += this.similar_text(first.substr(pos1 + max, firstLength - pos1 - max), second.substr(pos2 + max, secondLength - pos2 - max));
        }
    }

    return sum;
}

【讨论】:

  • 相似度是根据匹配字符返回的吗?它如何评估相似度
【解决方案7】:

fuzzyset - 为 javascript 设置的模糊字符串。 模糊集是一种数据结构,它对数据执行类似于全文搜索的操作,以确定可能的拼写错误和近似字符串匹配。请注意,这是一个 python 库的 javascript 端口。

【讨论】:

    【解决方案8】:

    在某种程度上,我喜欢Dice's coefficient 嵌入在string-similarity 模块中的想法。但我觉得只考虑二元组而不考虑它们的多重性会丢失一些重要数据。下面是一个也可以处理多重性的版本,我认为总体上是一个更简单的实现。我不尝试使用他们的 API,只提供一个函数,该函数在一些操作后比较两个字符串(删除非字母数字字符,小写所有内容,压缩但不删除空格),构建在一个比较它们而不进行操作的顶部.将其封装在他们的 API 中很容易,但我认为没有必要。

    const stringSimilarity = (a, b) =>
      _stringSimilarity (prep (a), prep (b))
    
    const _stringSimilarity = (a, b) => {
      const bg1 = bigrams (a)
      const bg2 = bigrams (b)
      const c1 = count (bg1)
      const c2 = count (bg2)
      const combined = uniq ([... bg1, ... bg2]) 
        .reduce ((t, k) => t + (Math .min (c1 [k] || 0, c2 [k] || 0)), 0)
      return 2 * combined / (bg1 .length + bg2 .length)
    }
    
    const prep = (str) => // TODO: unicode support?
      str .toLowerCase () .replace (/[^\w\s]/g, ' ') .replace (/\s+/g, ' ')
    
    const bigrams = (str) => 
      [...str] .slice (0, -1) .map ((c, i) => c + str [i + 1])
    
    const count = (xs) => 
      xs .reduce ((a, x) => ((a [x] = (a [x] || 0) + 1), a), {})
    
    const uniq = (xs) => 
      [... new Set (xs)]
    
    console .log (stringSimilarity (
      'foobar', 
      'Foobar'
    )) //=> 1
    
    console .log (stringSimilarity (
      "healed", 
      "sealed"
    ))//=> 0.8
    
    console .log (stringSimilarity (
      "Olive-green table for sale, in extremely good condition.",
      "For sale: table in very good  condition, olive green in colour."
    )) //=> 0.7787610619469026
    
    console .log (stringSimilarity (
      "Olive-green table for sale, in extremely good condition.",
      "For sale: green Subaru Impreza, 210,000 miles"
    )) //=> 0.38636363636363635
    
    console .log (stringSimilarity (
      "Olive-green table for sale, in extremely good condition.",
      "Wanted: mountain bike with at least 21 gears."
    )) //=> 0.1702127659574468
    
    console .log (stringSimilarity (
      "The rain in Spain falls mainly on the plain.",
      "The run in Spun falls munly on the plun.",
    )) //=> 0.7560975609756098
    
    console .log (stringSimilarity (
      "Fa la la la la, la la la la",
      "Fa la la la la, la la",
    )) //=> 0.8636363636363636
    
    console .log (stringSimilarity (
      "car crash",
      "carcrash",
    )) //=> 0.8
    
    console .log (stringSimilarity (
      "Now is the time for all good men to come to the aid of their party.",
      "Huh?",
    )) //=> 0
    .as-console-wrapper {max-height: 100% !important; top: 0}

    一些测试用例来自字符串相似性,其他的是我自己的。它们显示出与该软件包的一些显着差异,但没有什么不妥。我唯一要指出的是"car crash""carcrash" 之间的区别,字符串相似性认为是相同的,我报告的相似性为0.8。我的版本在所有橄榄绿测试用例中发现比字符串相似度更多的相似性,但由于这些在任何情况下都是相当任意的数字,我不确定它有多大的不同;他们当然以相同的相对顺序定位它们。

    【讨论】: