Naive Bayes using JavaScript

是 否 +2 论坛币

k人 参与回答

经管之家送您一份

应届毕业生专属福利!

求职就业群

赵安豆老师微信：zhaoandou666

经管之家联合CDA

送您一个全额奖学金名额~ !

立即领取

感谢您参与论坛问题回答

经管之家送您两个论坛币！

+2 论坛币

1. Methods
   
2. 
   
3. new NaiveBayes()
   
4. 
   
5. Constructor that takes no arguments.
   
6. 
   
7. Example
   
8. 
   
9. var nb = new NaiveBayes();
   
10. train(trainingSet, predictions)
    
11. 
    
12. Train the Naive Bayes model to the given training set and predictions
    
13. 
    
14. Arguments
    
15. 
    
16. trainingSet - A matrix of the training set.
    
17. trainingLabels - An array of value for each case in the training set.
    
18. Example
    
19. 
    
20. var cases = [[6,148,72,35,0,33.6,0.627,5],
    
21.              [1.50,85,66.5,29,0,26.6,0.351,31],
    
22.              [8,183,64,0,0,23.3,0.672,32],
    
23.              [0.5,89,65.5,23,94,28.1,0.167,21],
    
24.              [0,137,40,35,168,43.1,2.288,33]];
    
25. var predictions = [1, 0, 1, 0, 1];
    
26. 
    
27. nb.train(trainingSet, predictions);
    
28. predict(dataset)
    
29. 
    
30. Predict the values of the dataset.
    
31. 
    
32. Arguments
    
33. 
    
34. dataset - A matrix that contains the dataset.
    
35. Example
    
36. 
    
37. var dataset = [[6,148,72,35,0,33.6,0.627,5],
    
38.                [1.50,85,66.5,29,0,26.6,0.351,31]];
    
39. 
    
40. var ans = nb.predict(dataset);
    
41. export()
    
42. 
    
43. Exports the actual Naive Bayes model to an Javascript Object.
    
44. 
    
45. load(model)
    
46. 
    
47. Returns a new Naive Bayes Classifier with the given model.
    
48. 
    
49. Arguments
    
50. 
    
51. model - Javascript Object generated from export() function.
    
52. Authors
    
53. 
    
54. Jefferson Hernandez
    
55. License
    
56. 
    
57. MIT

复制代码

本帖隐藏的内容

naive-bayes-master.zip (6.34 KB)

2016-6-26 23:50:42 上传

扫码加我 拉你入群

请注明：姓名-公司-职位

以便审核进群资格，未注明则拒绝

分享0 收藏0 回帖

关键词：Javascript script naive scrip Using training matrix values

1. 'use strict';
   
2. 
   
3. var Matrix = require('ml-matrix');
   
4. var Stat = require('ml-stat');
   
5. 
   
6. module.exports.NaiveBayes = NaiveBayes;
   
7. module.exports.separateClasses = separateClasses;
   
8. 
   
9. /**
   
10. * Constructor for the Naive Bayes classifier, the parameters here is just for loading purposes.
    
11. *
    
12. * @param reload
    
13. * @param model
    
14. * @constructor
    
15. */
    
16. function NaiveBayes(reload, model) {
    
17.     if(reload) {
    
18.         this.means = model.means;
    
19.         this.calculateProbabilities = model.calculateProbabilities;
    
20.     }
    
21. }
    
22. 
    
23. /**
    
24. * Function that trains the classifier with a matrix that represents the training set and an array that
    
25. * represents the label of each row in the training set. the labels must be numbers between 0 to n-1 where
    
26. * n represents the number of classes.
    
27. *
    
28. * WARNING: in the case that one class, all the cases in one or more features have the same value, the
    
29. * Naive Bayes classifier will not work well.
    
30. * @param trainingSet
    
31. * @param trainingLabels
    
32. */
    
33. NaiveBayes.prototype.train = function (trainingSet, trainingLabels) {
    
34.     var C1 = Math.sqrt(2*Math.PI); // constant to precalculate the squared root
    
35.     if(!Matrix.isMatrix(trainingSet)) trainingSet = new Matrix(trainingSet);
    
36.     else trainingSet = trainingSet.clone();
    
37. 
    
38.     if(trainingSet.rows !== trainingLabels.length)
    
39.         throw new RangeError("the size of the training set and the training labels must be the same.");
    
40. 
    
41.     var separatedClasses = separateClasses(trainingSet, trainingLabels);
    
42.     var calculateProbabilities = new Array(separatedClasses.length);
    
43.     this.means = new Array(separatedClasses.length);
    
44.     for(var i = 0; i < separatedClasses.length; ++i) {
    
45.         var means = Stat.matrix.mean(separatedClasses[i]);
    
46.         var std = Stat.matrix.standardDeviation(separatedClasses[i], means);
    
47. 
    
48.         var logPriorProbability = Math.log(separatedClasses[i].rows / trainingSet.rows);
    
49.         calculateProbabilities[i] = new Array(means.length + 1);
    
50. 
    
51.         calculateProbabilities[i][0] = logPriorProbability;
    
52.         for(var j = 1; j < means.length + 1; ++j) {
    
53.             var currentStd = std[j - 1];
    
54.             calculateProbabilities[i][j] = [(1 / (C1 * currentStd)), -2*currentStd*currentStd];
    
55.         }
    
56. 
    
57.         this.means[i] = means;
    
58.     }
    
59. 
    
60.     this.calculateProbabilities = calculateProbabilities;
    
61. };
    
62. 
    
63. /**
    
64. * function that predicts each row of the dataset (must be a matrix).
    
65. *
    
66. * @param dataset
    
67. * @returns {Array}
    
68. */
    
69. NaiveBayes.prototype.predict = function (dataset) {
    
70.     if(dataset[0].length === this.calculateProbabilities[0].length)
    
71.         throw new RangeError('the dataset must have the same features as the training set');
    
72. 
    
73.     var predictions = new Array(dataset.length);
    
74. 
    
75.     for(var i = 0; i < predictions.length; ++i) {
    
76.         predictions[i] = getCurrentClass(dataset[i], this.means, this.calculateProbabilities);
    
77.     }
    
78. 
    
79.     return predictions;
    
80. };
    
81. 
    
82. /**
    
83. * Function the retrieves a prediction with one case.
    
84. *
    
85. * @param currentCase
    
86. * @param mean - Precalculated means of each class trained
    
87. * @param classes - Precalculated value of each class (Prior probability and probability function of each feature)
    
88. * @returns {number}
    
89. */
    
90. function getCurrentClass(currentCase, mean, classes) {
    
91.     var maxProbability = 0;
    
92.     var predictedClass = -1;
    
93. 
    
94.     // going through all precalculated values for the classes
    
95.     for(var i = 0; i < classes.length; ++i) {
    
96.         var currentProbability = classes[i][0]; // initialize with the prior probability
    
97.         for(var j = 1; j < classes[0][1].length + 1; ++j) {
    
98.             currentProbability += calculateLogProbability(currentCase[j - 1], mean[i][j - 1], classes[i][j][0], classes[i][j][1]);
    
99.         }
    
100. 
     
101.         currentProbability = Math.exp(currentProbability);
     
102.         if(currentProbability > maxProbability) {
     
103.             maxProbability = currentProbability;
     
104.             predictedClass = i;
     
105.         }
     
106.     }
     
107. 
     
108.     return predictedClass;
     
109. }
     
110. 
     
111. /**
     
112. * Function that export the NaiveBayes model.
     
113. * @returns {{modelName: string, means: *, calculateProbabilities: *}}
     
114. */
     
115. NaiveBayes.prototype.export = function () {
     
116.     return {
     
117.         modelName: "NaiveBayes",
     
118.         means: this.means,
     
119.         calculateProbabilities: this.calculateProbabilities
     
120.     };
     
121. };
     
122. 
     
123. /**
     
124. * Function that create a Naive Bayes classifier with the given model.
     
125. * @param model
     
126. * @returns {NaiveBayes}
     
127. */
     
128. NaiveBayes.load = function (model) {
     
129.     if(model.modelName !== 'NaiveBayes')
     
130.         throw new RangeError("The given model is invalid!");
     
131. 
     
132.     return new NaiveBayes(true, model);
     
133. };
     
134. 
     
135. /**
     
136. * function that retrieves the probability of the feature given the class.
     
137. * @param value - value of the feature.
     
138. * @param mean - mean of the feature for the given class.
     
139. * @param C1 - precalculated value of (1 / (sqrt(2*pi) * std)).
     
140. * @param C2 - precalculated value of (2 * std^2) for the denominator of the exponential.
     
141. * @returns {number}
     
142. */
     
143. function calculateLogProbability(value, mean, C1, C2) {
     
144.     var value = value - mean;
     
145.     return Math.log(C1 * Math.exp((value * value) / C2))
     
146. }
     
147. 
     
148. /**
     
149. * Function that retuns an array of matrices of the cases that belong to each class.
     
150. * @param X - dataset
     
151. * @param y - predictions
     
152. * @returns {Array}
     
153. */
     
154. function separateClasses(X, y) {
     
155.     var features = X.columns;
     
156. 
     
157.     var classes = 0;
     
158.     var totalPerClasses = new Array(100); // max upperbound of classes
     
159.     for (var i = 0; i < y.length; i++) {
     
160.         if(totalPerClasses[y[i]] === undefined) {
     
161.             totalPerClasses[y[i]] = 0;
     
162.             classes++;
     
163.         }
     
164.         totalPerClasses[y[i]]++;
     
165.     }
     
166.     var separatedClasses = new Array(classes);
     
167.     var currentIndex = new Array(classes);
     
168.     for(i = 0; i < classes; ++i) {
     
169.         separatedClasses[i] = new Matrix(totalPerClasses[i], features);
     
170.         currentIndex[i] = 0;
     
171.     }
     
172.     for(i = 0; i < X.rows; ++i) {
     
173.         separatedClasses[y[i]].setRow(currentIndex[y[i]], X.getRow(i));
     
174.         currentIndex[y[i]]++;
     
175.     }
     
176.     return separatedClasses;
     
177. }
     
178. Status API Training Shop Blog About
     
179. © 2016 GitHub, Inc. Terms Privacy Security Contact

复制代码

did you write these codes ? looks great !