文件列表:

A2B-AX-3y.csv (29217, 2020-07-21)
H2OAI.psd1 (9483, 2020-07-21)
H2OAI.psm1 (20580, 2020-07-21)
LICENSE (1072, 2020-07-21)

# H2O AI PowerShell Module This PowerShell Module is a simple wrapper for a series of POST requests to the [H2O AI](https://www.h2o.ai/) Open Source Platform Server. [](https://www.powershellgallery.com/packages/H2OAI) [](https://www.powershellgallery.com/packages/H2OAI) It is based off the work of Tome Tanasovski as detailed in his [blog post - H2o – Machine Learning with PowerShell](https://powertoe.wordpress.com/2017/10/23/h2o-machine-learning-with-powershell/) It contains three cmdlets of note for using the module; * Start-H2O (Start the H2O AI Server) * Stop-H2O (Stop the H2O AI Server) * Get-H2OPrediction (Get a Prediction using H2O AI) [Available in the PowerShell Gallery](https://www.powershellgallery.com/packages/H2OAI) [Associated Blogpost](https://blog.darrenjrobinson.com/h2o-ai-powershell-module/) ## Install Install direct from the PowerShell Gallery (Powershell 5.1 and above) ``` install-module -name H2OAI ``` ## H2O AI Algorithms [H2O AI Algorithms](http://docs.h2o.ai/h2o/latest-stable/h2o-docs/data-science.html) The H2O AI PowerShell Module will accept the following algorithms with Get-H2OPrediction. Your Training and Prediction data will need to be of the appropriate type for the algorithm to work. **Note** only GLM, GBM and DeepLearning have had any level of testing; - 'glm', 'gbm', '"glrm', 'aggregator', 'deeplearning', 'drf', 'isolationforest', 'kmeans', 'naivebayes', 'pca', 'targetencoder', 'word2vec' ## Prerequsites Java SE Runtime Environment I'm currently running [Java version "1.8.0_251"](https://www.oracle.com/java/technologies/javase/8u251-relnotes.html). ``` PS C:\Users\darrenjrobinson> java -version java version "1.8.0_251" Java(TM) SE Runtime Environment (build 1.8.0_251-b08) Java HotSpot(TM) 64-Bit Server VM (build 25.251-b08, mixed mode) ``` ## Download H2O AI [Download H2O AI](https://h2o-release.s3.amazonaws.com/h2o/rel-weierstrass/7/index.html) Extract to the local host (e.g. c:\h2o) **Note** H2O AI is currently a ~240Mb download (version 3.14.0.7). Uncompressed it is ~244Mb. ## Start H2OAI Import the H2O AI PowerShell Module and start H2O AI. ``` # Path to h2o.jar $dir = "C:\H2O\h2o-3.14.0.7" Start-H2o -H2oPath "$($dir)\h2o.jar" ``` # Import Training Data, Build a Model and make a Prediction Get-H2OPrediction is an all in one cmdlet to make using it super simple. Pass Get-H2OPrediction with; * a dataset * a model algorithm * a data split (defaults to 85% Train 15% Test) * data to make a prediction from and * the column to predict The default URL for the H2O AI Server is http://localhost:54321 **Note** The Predict Column name is case sensitive to what is in your dataset. If the dataset has the column heading as 'class' then you call Get-H2OPrediction with -predictColumn **Class** it will FAIL. ``` Get-H2OPrediction <# .SYNOPSIS Get an H2O Prediction for a dataset and a sample data request .DESCRIPTION Get an H2O Prediction for a dataset and a sample data request .PARAMETER url H2O URL default: "http://localhost:54321/3/{0}" .PARAMETER dataset H2O Dataset to build model from .PARAMETER predictData Data to build a prediction for .PARAMETER predictColumn Column to provide prediction on .PARAMETER modelAlgorithm Algorithm to build H2O model default: 'glm' .PARAMETER modelSplit Split of Model Dataset between Train and Test e.g ".85,.15" .EXAMPLE Get-H2OPrediction -url "http://localhost:54321/3/{0}" -dataset "c:\Data\dataSet.csv" -predictData = "c:\Data\predictDataSet.csv" -modelAlgorithm = 'glm' -modelSplit = ".85,.15" #> ``` ## Iris Example Below shows using Tome's the [Iris example](https://powertoe.wordpress.com/2017/10/23/h2o-machine-learning-with-powershell/) with the module. ``` # Default H2O AI Server running locally via Start-H2O $url = "http://localhost:54321/3/{0}" # Neural net algorithm for determining Iris type $modelAlgorithm = 'deeplearning' # Get Iris Training data and put on the local filesystem Invoke-RestMethod -Method Get 'https://raw.githubusercontent.com/DarrenCook/h2o/bk/datasets/iris_wheader.csv' | out-file ./iris_wheader.csv # Prediction Column $predictValues = 'class' # Data to make prediction from stored as a CSV on the local filesystem @" sepal_len, sepal_wid, petal_len, petal_wid 5.1,3.5,1.4,0.15 "@ | out-file -encoding ASCII ./iris_predict.csv # Send to H2O AI and get prediction $dataPath = (Get-ChildItem ./iris_predict.csv).DirectoryName $result = $null $result = Get-H2OPrediction -url $url -dataset "$($dataPath)/iris_wheader.csv" -predictData "$($dataPath)/iris_predict.csv" -modelAlgorithm $modelAlgorithm -modelSplit ".85,.15" -predictColumn $predictValues $result.prediction | Format-Table ``` ### Output ``` label data ----- ---- predict {0} Iris-setosa {0.999969054518561} Iris-versicolor {3.09454814387964E-05} Iris-virginica {2.15784441455808E-28} ``` ## Time Series Example Train a [Genearlised Data Model](http://docs.h2o.ai/h2o/latest-stable/h2o-docs/data-science/glm.html) with a time series dataset with a 85% to 15% split for Train and Test, and predict the next Close value. ### Example DataSet $sourceData | Select-Object -First 10 | Format-Table ``` Open High Close Low Volume Date ---- ---- ----- --- ------ ---- 3.68 3.8 3.74 3.68 50208 23-01-2017 3.74 3.75 3.75 3.66 47972 24-01-2017 3.75 3.8 3.8 3.73 33952 25-01-2017 3.8 3.8 3.79 3.77 32822 27-01-2017 3.8 3.8 3.73 3.68 21552 30-01-2017 3.75 3.75 3.65 3.6 50763 31-01-2017 3.69 3.69 3.64 3.61 59377 01-02-2017 3.64 3.66 3.55 3.51 120869 02-02-2017 3.64 3.66 3.49 3.49 75814 03-02-2017 3.49 3.54 3.44 3.43 86494 06-02-2017 ``` ``` # Linear Regression Model $modelAlgorithm = 'glm' # Time Series Data $dataCSV = "C:\Users\darrenjrobinson\Dropbox\Kloud\Projects\MLDoctaFileServer\data\A2B-AX-3y.csv" $sourceData = Import-Csv $dataCSV # Last Record as Prediction data $dataPredict = Import-Csv -Path $dataCSV | Select-Object -Last 1 | export-csv ./dataPredict.csv "Prediction Data" Import-Csv -Path $dataCSV | Select-Object -Last 1 # Predict Value $predictValues = 'Close' $result = $null $dataPath = (Get-ChildItem ./dataPredict.csv).DirectoryName $result = Get-H2OPrediction -url $url -dataset $dataCSV -predictData "$($dataPath)/dataPredict.csv" -modelAlgorithm $modelAlgorithm -modelSplit ".85,.15" -predictColumn $predictValues "Confidence: $($result.modelConfidence)" "Prediction: $($result.prediction.data)" ``` ### Output ``` Prediction Data Open : 1.46 High : 1.47 Close : 1.435 Low : 1.435 Volume : 17366 Date : 21-01-2020 Confidence: 0.000503095287004826 Prediction: 1.45785244828493 ``` # Stop H2O AI ``` Stop-H2O ``` ## Keep up to date * [Visit my blog](http://darrenjrobinson.com/) *  [Follow on Twitter](https://twitter.com/darrenjrobinson)

近期下载者：

相关文件：

评论：[我要评论] [举报此文件]

收藏者：