Merge pull request #281 from BrainJS/185-cross-validation-fixes

fix: Fix CrossValidate to have tests for when data too small

README.md

@@ -279,8 +279,9 @@ With multiple networks you can train in parallel like this:
 ### Cross Validation
 [Cross Validation](https://en.wikipedia.org/wiki/Cross-validation_(statistics)) can provide a less fragile way of training on larger data sets.  The brain.js api provides Cross Validation in this example:
 ```js
-const crossValidate = new CrossValidate(brain.NeuralNetwork, networkOptions);
-const stats = crossValidate.train(data, trainingOptions, k); //note k (or KFolds) is optional
+const crossValidate = new brain.CrossValidate(brain.NeuralNetwork, networkOptions);
+crossValidate.train(data, trainingOptions, k); //note k (or KFolds) is optional
+const json = crossValidate.toJSON(); // all stats in json as well as neural networks
 const net = crossValidate.toNeuralNetwork();
 
 

bower.json

@@ -31,5 +31,5 @@
     "node_modules",
     "test"
   ],
-  "version": "1.4.1"
+  "version": "1.4.2"
 }

browser.js

@@ -6,7 +6,7 @@
  *   license: MIT (http://opensource.org/licenses/MIT)
  *   author: Heather Arthur <[email protected]>
  *   homepage: https://github.com/brainjs/brain.js#readme
- *   version: 1.4.1
+ *   version: 1.4.2
  *
  * acorn:
  *   license: MIT (http://opensource.org/licenses/MIT)
@@ -214,8 +214,13 @@ var CrossValidate = function () {
 
   }, {
     key: "train",
-    value: function train(data, trainOpts, k) {
-      k = k || 4;
+    value: function train(data) {
+      var trainOpts = arguments.length > 1 && arguments[1] !== undefined ? arguments[1] : {};
+      var k = arguments.length > 2 && arguments[2] !== undefined ? arguments[2] : 4;
+
+      if (data.length <= k) {
+        throw new Error("Training set size is too small for " + data.length + " k folds of " + k);
+      }
       var size = data.length / k;
 
       if (data.constructor === Array) {
@@ -1946,8 +1951,8 @@ var NeuralNetwork = function () {
           falseNeg: falseNeg,
           falsePos: falsePos,
           total: data.length,
-          precision: truePos / (truePos + falsePos),
-          recall: truePos / (truePos + falseNeg),
+          precision: truePos > 0 ? truePos / (truePos + falsePos) : 0,
+          recall: truePos > 0 ? truePos / (truePos + falseNeg) : 0,
           accuracy: (trueNeg + truePos) / data.length
         });
       }

examples-typescript/cross-validate.ts

@@ -8,23 +8,11 @@ const trainingData = [
   { input: [1, 1], output: [0] },
   { input: [1, 0], output: [1] },
 
-  // xor repeats
+  // repeat xor data to have enough to train with
   { input: [0, 1], output: [1] },
   { input: [0, 0], output: [0] },
   { input: [1, 1], output: [0] },
-  { input: [1, 0], output: [1] },
-
-  // xor repeats
-  { input: [0, 1], output: [1] },
-  { input: [0, 0], output: [0] },
-  { input: [1, 1], output: [0] },
-  { input: [1, 0], output: [1] },
-
-  // xor repeats
-  { input: [0, 1], output: [1] },
-  { input: [0, 0], output: [0] },
-  { input: [1, 1], output: [0] },
-  { input: [1, 0], output: [1] },
+  { input: [1, 0], output: [1] }
 ];
 
 const netOptions = {

examples/cross-validate.js

@@ -8,23 +8,11 @@ const trainingData = [
   { input: [1, 1], output: [0] },
   { input: [1, 0], output: [1] },
 
-  // xor repeats
+  // repeat xor data to have enough to train with
   { input: [0, 1], output: [1] },
   { input: [0, 0], output: [0] },
   { input: [1, 1], output: [0] },
-  { input: [1, 0], output: [1] },
-
-  // xor repeats
-  { input: [0, 1], output: [1] },
-  { input: [0, 0], output: [0] },
-  { input: [1, 1], output: [0] },
-  { input: [1, 0], output: [1] },
-
-  // xor repeats
-  { input: [0, 1], output: [1] },
-  { input: [0, 0], output: [0] },
-  { input: [1, 1], output: [0] },
-  { input: [1, 0], output: [1] },
+  { input: [1, 0], output: [1] }
 ];
 
 const netOptions = {

package.json

@@ -1,7 +1,7 @@
 {
   "name": "brain.js",
   "description": "Neural network library",
-  "version": "1.4.1",
+  "version": "1.4.2",
   "author": "Heather Arthur <[email protected]>",
   "repository": {
     "type": "git",

src/cross-validate.js

@@ -78,8 +78,10 @@ export default class CrossValidate {
    *  }
    * }
    */
-  train(data, trainOpts, k) {
-    k = k || 4;
+  train(data, trainOpts = {}, k = 4) {
+    if (data.length <= k) {
+      throw new Error(`Training set size is too small for ${ data.length } k folds of ${ k }`);
+    }
     let size = data.length / k;
 
     if (data.constructor === Array) {

src/neural-network.js

@@ -805,8 +805,8 @@ export default class NeuralNetwork {
         falseNeg: falseNeg,
         falsePos: falsePos,
         total: data.length,
-        precision: truePos / (truePos + falsePos),
-        recall: truePos / (truePos + falseNeg),
+        precision: truePos > 0 ? truePos / (truePos + falsePos) : 0,
+        recall: truePos > 0 ? truePos / (truePos + falseNeg) : 0,
         accuracy: (trueNeg + truePos) / data.length
       });
     }

test/base/cross-validation.js

@@ -0,0 +1,42 @@
+import assert from 'assert';
+import brain from '../../src';
+import CrossValidate from '../../src/cross-validate';
+
+describe('CrossValidation', () => {
+  describe('simple xor example', () => {
+    it('throws exception when training set is too small', () => {
+      const xorTrainingData = [
+        { input: [0, 1], output: [1] },
+        { input: [0, 0], output: [0] },
+        { input: [1, 1], output: [0] },
+        { input: [1, 0], output: [1] }
+      ];
+      const net = new CrossValidate(brain.NeuralNetwork);
+      assert.throws(() => {
+        net.train(xorTrainingData);
+      });
+    });
+    it('handles training and outputs values that are all numbers', () => {
+      const xorTrainingData = [
+        { input: [0, 1], output: [1] },
+        { input: [0, 0], output: [0] },
+        { input: [1, 1], output: [0] },
+        { input: [1, 0], output: [1] },
+
+        { input: [0, 1], output: [1] },
+        { input: [0, 0], output: [0] },
+        { input: [1, 1], output: [0] },
+        { input: [1, 0], output: [1] }
+      ];
+      const net = new CrossValidate(brain.NeuralNetwork);
+      net.train(xorTrainingData);
+      const json = net.toJSON();
+      for (let p in json.avgs) {
+        assert(json.avgs[p] >= 0);
+      }
+      for (let p in json.stats) {
+        assert(json.stats[p] >= 0);
+      }
+    });
+  });
+});