Question

What’s the best and most efficient way to count keywords in JavaScript? Basically, I’d like to take a string and get the top N words or phrases that occur in the string, mainly for the use of suggesting tags. I’m looking more for conceptual hints or links to real-life examples than actual code, but I certainly wouldn’t mind if you’d like to share code as well. If there are particular functions that would help, I’d also appreciate that.

Right now I think I’m at using the split() function to separate the string by spaces and then cleaning punctuation out with a regular expression. I’d also want it to be case-insensitive.

Answer 1

Cut, paste + execute demo:

var text = 'Text to be examined to determine which n words are used the most';  // Find 'em! var wordRegExp = /\w+(?:'\w{1,2})?/g; var words = {}; var matches; while ((matches = wordRegExp.exec(text)) != null) {     var word = matches[0].toLowerCase();     if (typeof words[word] == 'undefined')     {         words[word] = 1;     }     else     {         words[word]++;     } }  // Sort 'em! var wordList = []; for (var word in words) {     if (words.hasOwnProperty(word))     {         wordList.push([word, words[word]]);     } } wordList.sort(function(a, b) { return b[1] - a[1]; });  // Come back any time, straaanger! var n = 10; var message = ['The top ' + n + ' words are:']; for (var i = 0; i < n; i++) {     message.push(wordList[i][0] + ' - ' + wordList[i][1] + ' occurance' +                  (wordList[i][1] == 1 ? '' : 's')); } alert(message.join('\n')); 

Reusable function:

function getTopNWords(text, n) {     var wordRegExp = /\w+(?:'\w{1,2})?/g;     var words = {};     var matches;     while ((matches = wordRegExp.exec(text)) != null)     {         var word = matches[0].toLowerCase();         if (typeof words[word] == 'undefined')         {             words[word] = 1;         }         else         {             words[word]++;         }     }      var wordList = [];     for (var word in words)     {         if (words.hasOwnProperty(word))         {             wordList.push([word, words[word]]);         }     }     wordList.sort(function(a, b) { return b[1] - a[1]; });      var topWords = [];     for (var i = 0; i < n; i++)     {         topWords.push(wordList[i][0]);     }     return topWords; }