【JavaScript】Classification and Regression Trees using JavaScript

相似文件 换一批

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

CART (Classification and regression trees)

Decision trees using CART implementation.

Installation

npm install ml-cart

DocumentationLicense

MIT

本帖隐藏的内容

decision-tree-cart-master.zip (1.16 MB)

2017-4-16 04:24:42 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3 回帖

关键词：Javascript regression regressio regress script

1. 'use strict';
   
2. 
   
3. var Tree = require('./TreeNode');
   
4. var Matrix = require('ml-matrix');
   
5. 
   
6. class DecisionTreeClassifier {
   
7. 
   
8.     /**
   
9.      * Create new Decision Tree Classifier with CART implementation with the given options
   
10.      * @param {object} options
    
11.      * @param {string} [options.gainFunction="gini"] - gain function to get the best split, "gini" the only one supported.
    
12.      * @param {string} [options.splitFunction] - given two integers from a split feature, get the value to split, "mean" the only one supported.
    
13.      * @param {number} [options.minNumSamples] - minimum number of samples to create a leaf node to decide a class. Default 3.
    
14.      * @param {number} [options.maxDepth] - Max depth of the tree. Default Infinity.
    
15.      * @param {object} model - for load purposes.
    
16.      * @constructor
    
17.      */
    
18.     constructor(options, model) {
    
19.         if (options === true) {
    
20.             this.options = model.options;
    
21.             this.root = new Tree(model.options);
    
22.             this.root.setNodeParameters(model.root);
    
23.         } else {
    
24.             if (options === undefined) options = {};
    
25.             if (options.gainFunction === undefined) options.gainFunction = 'gini';
    
26.             if (options.splitFunction === undefined) options.splitFunction = 'mean';
    
27.             if (options.minNumSamples === undefined) options.minNumSamples = 3;
    
28.             if (options.maxDepth === undefined) options.maxDepth = Infinity;
    
29. 
    
30.             options.kind = 'classifier';
    
31.             this.options = options;
    
32.         }
    
33.     }
    
34. 
    
35.     /**
    
36.      * Train the decision tree with the given training set and labels.
    
37.      * @param {Matrix} trainingSet
    
38.      * @param {Array} trainingLabels
    
39.      */
    
40.     train(trainingSet, trainingLabels) {
    
41.         this.root = new Tree(this.options);
    
42.         if (!Matrix.isMatrix(trainingSet)) trainingSet = new Matrix(trainingSet);
    
43.         this.root.train(trainingSet, trainingLabels, 0, null);
    
44.     }
    
45. 
    
46.     /**
    
47.      * Predicts the output given the matrix to predict.
    
48.      * @param {Matrix} toPredict
    
49.      * @return {Array} predictions
    
50.      */
    
51.     predict(toPredict) {
    
52.         var predictions = new Array(toPredict.length);
    
53. 
    
54.         for (var i = 0; i < toPredict.length; ++i) {
    
55.             predictions[i] = this.root.classify(toPredict[i]).maxRowIndex(0)[1];
    
56.         }
    
57. 
    
58.         return predictions;
    
59.     }
    
60. 
    
61.     /**
    
62.      * Export the current model to JSON.
    
63.      * @return {object} - Current model.
    
64.      */
    
65.     toJSON() {
    
66.         var toSave = {
    
67.             options: this.options,
    
68.             root: {},
    
69.             name: 'DTClassifier'
    
70.         };
    
71. 
    
72.         toSave.root = this.root;
    
73.         return toSave;
    
74.     }
    
75. 
    
76.     /**
    
77.      * Load a Decision tree classifier with the given model.
    
78.      * @param {object} model
    
79.      * @return {DecisionTreeClassifier}
    
80.      */
    
81.     static load(model) {
    
82.         if (model.name !== 'DTClassifier') {
    
83.             throw new RangeError('Invalid model: ' + model.name);
    
84.         }
    
85. 
    
86.         return new DecisionTreeClassifier(true, model);
    
87.     }
    
88. }
    
89. 
    
90. module.exports = DecisionTreeClassifier;

复制代码

1. 'use strict';
   
2. 
   
3. var Tree = require('./TreeNode');
   
4. var Matrix = require('ml-matrix');
   
5. 
   
6. class DecisionTreeRegression {
   
7. 
   
8.     /**
   
9.      * Create new Decision Tree Regression with CART implementation with the given options.
   
10.      * @param {object} options
    
11.      * @param {string} [options.gainFunction="regression"] - gain function to get the best split, "regression" the only one supported.
    
12.      * @param {string} [options.splitFunction] - given two integers from a split feature, get the value to split, "mean" the only one supported.
    
13.      * @param {number} [options.minNumSamples] - minimum number of samples to create a leaf node to decide a class. Default 3.
    
14.      * @param {number} [options.maxDepth] - Max depth of the tree. Default Infinity.
    
15.      * @param {object} model - for load purposes.
    
16.      */
    
17.     constructor(options, model) {
    
18.         if (options === true) {
    
19.             this.options = model.options;
    
20.             this.root = new Tree(model.options);
    
21.             this.root.setNodeParameters(model.root);
    
22.         } else {
    
23.             if (options === undefined) options = {};
    
24.             if (options.gainFunction === undefined) options.gainFunction = 'regression';
    
25.             if (options.splitFunction === undefined) options.splitFunction = 'mean';
    
26.             if (options.minNumSamples === undefined) options.minNumSamples = 3;
    
27.             if (options.maxDepth === undefined) options.maxDepth = Infinity;
    
28. 
    
29.             options.kind = 'regression';
    
30.             this.options = options;
    
31.         }
    
32.     }
    
33. 
    
34.     /**
    
35.      * Train the decision tree with the given training set and values.
    
36.      * @param {Matrix} trainingSet
    
37.      * @param {Array} trainingValues
    
38.      */
    
39.     train(trainingSet, trainingValues) {
    
40.         this.root = new Tree(this.options);
    
41.         if (trainingSet[0].length === undefined) trainingSet = Matrix.columnVector(trainingSet);
    
42.         if (!Matrix.isMatrix(trainingSet)) trainingSet = new Matrix(trainingSet);
    
43.         this.root.train(trainingSet, trainingValues, 0);
    
44.     }
    
45. 
    
46.     /**
    
47.      * Predicts the values given the matrix to predict.
    
48.      * @param {Matrix} toPredict
    
49.      * @return {Array} predictions
    
50.      */
    
51.     predict(toPredict) {
    
52.         if (toPredict[0].length === undefined) toPredict = Matrix.columnVector(toPredict);
    
53.         var predictions = new Array(toPredict.length);
    
54. 
    
55.         for (var i = 0; i < toPredict.length; ++i) {
    
56.             predictions[i] = this.root.classify(toPredict[i]);
    
57.         }
    
58. 
    
59.         return predictions;
    
60.     }
    
61. 
    
62.     /**
    
63.      * Export the current model to JSON.
    
64.      * @return {object} - Current model.
    
65.      */
    
66.     toJSON() {
    
67.         var toSave = {
    
68.             options: this.options,
    
69.             root: {},
    
70.             name: 'DTRegression'
    
71.         };
    
72. 
    
73.         toSave.root = this.root;
    
74.         return toSave;
    
75.     }
    
76. 
    
77.     /**
    
78.      * Load a Decision tree regression with the given model.
    
79.      * @param {object} model
    
80.      * @return {DecisionTreeRegression}
    
81.      */
    
82.     static load(model) {
    
83.         if (model.name !== 'DTRegression') {
    
84.             throw new RangeError('Invalid model:' + model.name);
    
85.         }
    
86. 
    
87.         return new DecisionTreeRegression(true, model);
    
88.     }
    
89. }
    
90. 
    
91. module.exports = DecisionTreeRegression;

复制代码

好东西啊，学习一下

好东西啊，学习一下

thanks for sharing