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

Also upgrade to use newer es6 syntax for defining optional properties.
Upgrade examples to be more straightforward when using CrossValidate.
Fix values from being set to NaN when training with smaller data in CrossValidate.

Author: robertleeplummerjr

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 <fayearthur@gmail.com>
  *   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
         });
       }

browser.min.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 = {

dist/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 = {