[
https://issues.apache.org/jira/browse/FLINK-2259?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15286356#comment-15286356
]
ASF GitHub Bot commented on FLINK-2259:
---------------------------------------
Github user tillrohrmann commented on a diff in the pull request:
https://github.com/apache/flink/pull/1898#discussion_r63497505
--- Diff: docs/apis/batch/libs/ml/cross_validation.md ---
@@ -0,0 +1,175 @@
+---
+mathjax: include
+title: Cross Validation
+
+# Sub navigation
+sub-nav-group: batch
+sub-nav-parent: flinkml
+sub-nav-title: Cross Validation
+---
+<!--
+Licensed to the Apache Software Foundation (ASF) under one
+or more contributor license agreements. See the NOTICE file
+distributed with this work for additional information
+regarding copyright ownership. The ASF licenses this file
+to you under the Apache License, Version 2.0 (the
+"License"); you may not use this file except in compliance
+with the License. You may obtain a copy of the License at
+
+ http://www.apache.org/licenses/LICENSE-2.0
+
+Unless required by applicable law or agreed to in writing,
+software distributed under the License is distributed on an
+"AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
+KIND, either express or implied. See the License for the
+specific language governing permissions and limitations
+under the License.
+-->
+
+* This will be replaced by the TOC
+{:toc}
+
+## Description
+
+ A prevalent problem when utilizing machine learning algorithms is
*overfitting*, or when an algorithm "memorizes" the training data but does a
poor job extrapolating to out of sample cases. A common method for dealing with
the overfitting problem is to hold back some subset of data from the original
training algorithm and then measure the fit algorithm's performance on this
hold-out set. This is commonly known as *cross validation*. A model is trained
on one subset of data and then *validated* on another set of data.
+
+## Cross Validation Strategies
+
+There are several strategies for holding out data. FlinkML has convenience
methods for
+- Train-Test Splits
+- Train-Test-Holdout Splits
+- K-Fold Splits
+- Multi-Random Splits
+
+### Train-Test Splits
+
+The simplest method of splitting is the `trainTestSplit`. This split takes
a DataSet and a parameter *fraction*. The *fraction* indicates the portion of
the DataSet that should be allocated to the training set. This split also takes
two additional optional parameters, *precise* and *seed*.
+
+By default, the Split is done by randomly deciding weather or not an
observation is assigned to the training DataSet with probability = *fraction*.
When *precise* is `true` however, additional steps are taken to ensure the
training set is as close as possible to the length of the DataSet $\cdot$
*fraction*.
--- End diff --
typo: weather --> whether
> Support training Estimators using a (train, validation, test) split of the
> available data
> -----------------------------------------------------------------------------------------
>
> Key: FLINK-2259
> URL: https://issues.apache.org/jira/browse/FLINK-2259
> Project: Flink
> Issue Type: New Feature
> Components: Machine Learning Library
> Reporter: Theodore Vasiloudis
> Assignee: Trevor Grant
> Priority: Minor
> Labels: ML
>
> When there is an abundance of data available, a good way to train models is
> to split the available data into 3 parts: Train, Validation and Test.
> We use the Train data to train the model, the Validation part is used to
> estimate the test error and select hyperparameters, and the Test is used to
> evaluate the performance of the model, and assess its generalization [1]
> This is a common approach when training Artificial Neural Networks, and a
> good strategy to choose in data-rich environments. Therefore we should have
> some support of this data-analysis process in our Estimators.
> [1] Friedman, Jerome, Trevor Hastie, and Robert Tibshirani. The elements of
> statistical learning. Vol. 1. Springer, Berlin: Springer series in
> statistics, 2001.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
