| |
| <!DOCTYPE HTML> |
| <html lang="" > |
| <head> |
| <meta charset="UTF-8"> |
| <meta content="text/html; charset=utf-8" http-equiv="Content-Type"> |
| <title>XGBoost ยท Hivemall User Manual</title> |
| <meta http-equiv="X-UA-Compatible" content="IE=edge" /> |
| <meta name="description" content=""> |
| <meta name="generator" content="GitBook 3.2.3"> |
| |
| |
| |
| |
| <link rel="stylesheet" href="../gitbook/style.css"> |
| |
| |
| |
| |
| <link rel="stylesheet" href="../gitbook/gitbook-plugin-splitter/splitter.css"> |
| |
| |
| |
| <link rel="stylesheet" href="../gitbook/gitbook-plugin-etoc/plugin.css"> |
| |
| |
| |
| <link rel="stylesheet" href="../gitbook/gitbook-plugin-callouts/plugin.css"> |
| |
| |
| |
| <link rel="stylesheet" href="../gitbook/gitbook-plugin-toggle-chapters/toggle.css"> |
| |
| |
| |
| <link rel="stylesheet" href="../gitbook/gitbook-plugin-codeblock-filename/block.css"> |
| |
| |
| |
| <link rel="stylesheet" href="../gitbook/gitbook-plugin-expandable-chapters/expandable-chapters.css"> |
| |
| |
| |
| <link rel="stylesheet" href="../gitbook/gitbook-plugin-multipart/multipart.css"> |
| |
| |
| |
| <link rel="stylesheet" href="../gitbook/gitbook-plugin-katex/katex.min.css"> |
| |
| |
| |
| <link rel="stylesheet" href="../gitbook/gitbook-plugin-emphasize/plugin.css"> |
| |
| |
| |
| <link rel="stylesheet" href="../gitbook/gitbook-plugin-highlight/website.css"> |
| |
| |
| |
| <link rel="stylesheet" href="../gitbook/gitbook-plugin-search/search.css"> |
| |
| |
| |
| <link rel="stylesheet" href="../gitbook/gitbook-plugin-fontsettings/website.css"> |
| |
| |
| |
| <link rel="stylesheet" href="../gitbook/gitbook-plugin-theme-api/theme-api.css"> |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| <meta name="HandheldFriendly" content="true"/> |
| <meta name="viewport" content="width=device-width, initial-scale=1, user-scalable=no"> |
| <meta name="apple-mobile-web-app-capable" content="yes"> |
| <meta name="apple-mobile-web-app-status-bar-style" content="black"> |
| <link rel="apple-touch-icon-precomposed" sizes="152x152" href="../gitbook/images/apple-touch-icon-precomposed-152.png"> |
| <link rel="shortcut icon" href="../gitbook/images/favicon.ico" type="image/x-icon"> |
| |
| |
| <link rel="next" href="kdd2010a.html" /> |
| |
| |
| <link rel="prev" href="news20_rf.html" /> |
| |
| |
| </head> |
| <body> |
| |
| <div class="book"> |
| <div class="book-summary"> |
| |
| |
| <div id="book-search-input" role="search"> |
| <input type="text" placeholder="Type to search" /> |
| </div> |
| |
| |
| <nav role="navigation"> |
| |
| |
| |
| <ul class="summary"> |
| |
| |
| |
| |
| <li> |
| <a href="https://hivemall.incubator.apache.org/" target="_blank" class="custom-link"><i class="fa fa-home"></i> Home</a> |
| </li> |
| |
| |
| |
| |
| <li class="divider"></li> |
| |
| |
| |
| |
| <li class="header">TABLE OF CONTENTS</li> |
| |
| |
| |
| <li class="chapter " data-level="1.1" data-path="../"> |
| |
| <a href="../"> |
| |
| |
| <b>1.1.</b> |
| |
| Introduction |
| |
| </a> |
| |
| |
| |
| </li> |
| |
| <li class="chapter " data-level="1.2" data-path="../getting_started/"> |
| |
| <a href="../getting_started/"> |
| |
| |
| <b>1.2.</b> |
| |
| Getting Started |
| |
| </a> |
| |
| |
| |
| <ul class="articles"> |
| |
| |
| <li class="chapter " data-level="1.2.1" data-path="../getting_started/installation.html"> |
| |
| <a href="../getting_started/installation.html"> |
| |
| |
| <b>1.2.1.</b> |
| |
| Installation |
| |
| </a> |
| |
| |
| |
| </li> |
| |
| <li class="chapter " data-level="1.2.2" data-path="../getting_started/permanent-functions.html"> |
| |
| <a href="../getting_started/permanent-functions.html"> |
| |
| |
| <b>1.2.2.</b> |
| |
| Install as permanent functions |
| |
| </a> |
| |
| |
| |
| </li> |
| |
| <li class="chapter " data-level="1.2.3" data-path="../getting_started/input-format.html"> |
| |
| <a href="../getting_started/input-format.html"> |
| |
| |
| <b>1.2.3.</b> |
| |
| Input Format |
| |
| </a> |
| |
| |
| |
| </li> |
| |
| |
| </ul> |
| |
| </li> |
| |
| <li class="chapter " data-level="1.3" data-path="../misc/funcs.html"> |
| |
| <a href="../misc/funcs.html"> |
| |
| |
| <b>1.3.</b> |
| |
| List of Functions |
| |
| </a> |
| |
| |
| |
| </li> |
| |
| <li class="chapter " data-level="1.4" data-path="../tips/"> |
| |
| <a href="../tips/"> |
| |
| |
| <b>1.4.</b> |
| |
| Tips for Effective Hivemall |
| |
| </a> |
| |
| |
| |
| <ul class="articles"> |
| |
| |
| <li class="chapter " data-level="1.4.1" data-path="../tips/addbias.html"> |
| |
| <a href="../tips/addbias.html"> |
| |
| |
| <b>1.4.1.</b> |
| |
| Explicit add_bias() for better prediction |
| |
| </a> |
| |
| |
| |
| </li> |
| |
| <li class="chapter " data-level="1.4.2" data-path="../tips/rand_amplify.html"> |
| |
| <a href="../tips/rand_amplify.html"> |
| |
| |
| <b>1.4.2.</b> |
| |
| Use rand_amplify() to better prediction results |
| |
| </a> |
| |
| |
| |
| </li> |
| |
| <li class="chapter " data-level="1.4.3" data-path="../tips/rt_prediction.html"> |
| |
| <a href="../tips/rt_prediction.html"> |
| |
| |
| <b>1.4.3.</b> |
| |
| Real-time prediction on RDBMS |
| |
| </a> |
| |
| |
| |
| </li> |
| |
| <li class="chapter " data-level="1.4.4" data-path="../tips/ensemble_learning.html"> |
| |
| <a href="../tips/ensemble_learning.html"> |
| |
| |
| <b>1.4.4.</b> |
| |
| Ensemble learning for stable prediction |
| |
| </a> |
| |
| |
| |
| </li> |
| |
| <li class="chapter " data-level="1.4.5" data-path="../tips/mixserver.html"> |
| |
| <a href="../tips/mixserver.html"> |
| |
| |
| <b>1.4.5.</b> |
| |
| Mixing models for a better prediction convergence (MIX server) |
| |
| </a> |
| |
| |
| |
| </li> |
| |
| <li class="chapter " data-level="1.4.6" data-path="../tips/emr.html"> |
| |
| <a href="../tips/emr.html"> |
| |
| |
| <b>1.4.6.</b> |
| |
| Run Hivemall on Amazon Elastic MapReduce |
| |
| </a> |
| |
| |
| |
| </li> |
| |
| |
| </ul> |
| |
| </li> |
| |
| <li class="chapter " data-level="1.5" data-path="../tips/general_tips.html"> |
| |
| <a href="../tips/general_tips.html"> |
| |
| |
| <b>1.5.</b> |
| |
| General Hive/Hadoop Tips |
| |
| </a> |
| |
| |
| |
| <ul class="articles"> |
| |
| |
| <li class="chapter " data-level="1.5.1" data-path="../tips/rowid.html"> |
| |
| <a href="../tips/rowid.html"> |
| |
| |
| <b>1.5.1.</b> |
| |
| Adding rowid for each row |
| |
| </a> |
| |
| |
| |
| </li> |
| |
| <li class="chapter " data-level="1.5.2" data-path="../tips/hadoop_tuning.html"> |
| |
| <a href="../tips/hadoop_tuning.html"> |
| |
| |
| <b>1.5.2.</b> |
| |
| Hadoop tuning for Hivemall |
| |
| </a> |
| |
| |
| |
| </li> |
| |
| |
| </ul> |
| |
| </li> |
| |
| <li class="chapter " data-level="1.6" data-path="../troubleshooting/"> |
| |
| <a href="../troubleshooting/"> |
| |
| |
| <b>1.6.</b> |
| |
| Troubleshooting |
| |
| </a> |
| |
| |
| |
| <ul class="articles"> |
| |
| |
| <li class="chapter " data-level="1.6.1" data-path="../troubleshooting/oom.html"> |
| |
| <a href="../troubleshooting/oom.html"> |
| |
| |
| <b>1.6.1.</b> |
| |
| OutOfMemoryError in training |
| |
| </a> |
| |
| |
| |
| </li> |
| |
| <li class="chapter " data-level="1.6.2" data-path="../troubleshooting/mapjoin_task_error.html"> |
| |
| <a href="../troubleshooting/mapjoin_task_error.html"> |
| |
| |
| <b>1.6.2.</b> |
| |
| SemanticException generate map join task error: Cannot serialize object |
| |
| </a> |
| |
| |
| |
| </li> |
| |
| <li class="chapter " data-level="1.6.3" data-path="../troubleshooting/asterisk.html"> |
| |
| <a href="../troubleshooting/asterisk.html"> |
| |
| |
| <b>1.6.3.</b> |
| |
| Asterisk argument for UDTF does not work |
| |
| </a> |
| |
| |
| |
| </li> |
| |
| <li class="chapter " data-level="1.6.4" data-path="../troubleshooting/num_mappers.html"> |
| |
| <a href="../troubleshooting/num_mappers.html"> |
| |
| |
| <b>1.6.4.</b> |
| |
| The number of mappers is less than input splits in Hadoop 2.x |
| |
| </a> |
| |
| |
| |
| </li> |
| |
| <li class="chapter " data-level="1.6.5" data-path="../troubleshooting/mapjoin_classcastex.html"> |
| |
| <a href="../troubleshooting/mapjoin_classcastex.html"> |
| |
| |
| <b>1.6.5.</b> |
| |
| Map-side join causes ClassCastException on Tez |
| |
| </a> |
| |
| |
| |
| </li> |
| |
| |
| </ul> |
| |
| </li> |
| |
| |
| |
| |
| <li class="header">Part II - Generic Features</li> |
| |
| |
| |
| <li class="chapter " data-level="2.1" data-path="../misc/generic_funcs.html"> |
| |
| <a href="../misc/generic_funcs.html"> |
| |
| |
| <b>2.1.</b> |
| |
| List of Generic Hivemall Functions |
| |
| </a> |
| |
| |
| |
| </li> |
| |
| <li class="chapter " data-level="2.2" data-path="../misc/topk.html"> |
| |
| <a href="../misc/topk.html"> |
| |
| |
| <b>2.2.</b> |
| |
| Efficient Top-K Query Processing |
| |
| </a> |
| |
| |
| |
| </li> |
| |
| <li class="chapter " data-level="2.3" data-path="../misc/tokenizer.html"> |
| |
| <a href="../misc/tokenizer.html"> |
| |
| |
| <b>2.3.</b> |
| |
| Text Tokenizer |
| |
| </a> |
| |
| |
| |
| </li> |
| |
| <li class="chapter " data-level="2.4" data-path="../misc/approx.html"> |
| |
| <a href="../misc/approx.html"> |
| |
| |
| <b>2.4.</b> |
| |
| Approximate Aggregate Functions |
| |
| </a> |
| |
| |
| |
| </li> |
| |
| |
| |
| |
| <li class="header">Part III - Feature Engineering</li> |
| |
| |
| |
| <li class="chapter " data-level="3.1" data-path="../ft_engineering/scaling.html"> |
| |
| <a href="../ft_engineering/scaling.html"> |
| |
| |
| <b>3.1.</b> |
| |
| Feature Scaling |
| |
| </a> |
| |
| |
| |
| </li> |
| |
| <li class="chapter " data-level="3.2" data-path="../ft_engineering/hashing.html"> |
| |
| <a href="../ft_engineering/hashing.html"> |
| |
| |
| <b>3.2.</b> |
| |
| Feature Hashing |
| |
| </a> |
| |
| |
| |
| </li> |
| |
| <li class="chapter " data-level="3.3" data-path="../ft_engineering/selection.html"> |
| |
| <a href="../ft_engineering/selection.html"> |
| |
| |
| <b>3.3.</b> |
| |
| Feature Selection |
| |
| </a> |
| |
| |
| |
| </li> |
| |
| <li class="chapter " data-level="3.4" data-path="../ft_engineering/binning.html"> |
| |
| <a href="../ft_engineering/binning.html"> |
| |
| |
| <b>3.4.</b> |
| |
| Feature Binning |
| |
| </a> |
| |
| |
| |
| </li> |
| |
| <li class="chapter " data-level="3.5" data-path="../ft_engineering/pairing.html"> |
| |
| <a href="../ft_engineering/pairing.html"> |
| |
| |
| <b>3.5.</b> |
| |
| Feature Paring |
| |
| </a> |
| |
| |
| |
| <ul class="articles"> |
| |
| |
| <li class="chapter " data-level="3.5.1" data-path="../ft_engineering/polynomial.html"> |
| |
| <a href="../ft_engineering/polynomial.html"> |
| |
| |
| <b>3.5.1.</b> |
| |
| Polynomial features |
| |
| </a> |
| |
| |
| |
| </li> |
| |
| |
| </ul> |
| |
| </li> |
| |
| <li class="chapter " data-level="3.6" data-path="../ft_engineering/ft_trans.html"> |
| |
| <a href="../ft_engineering/ft_trans.html"> |
| |
| |
| <b>3.6.</b> |
| |
| Feature Transformation |
| |
| </a> |
| |
| |
| |
| <ul class="articles"> |
| |
| |
| <li class="chapter " data-level="3.6.1" data-path="../ft_engineering/vectorization.html"> |
| |
| <a href="../ft_engineering/vectorization.html"> |
| |
| |
| <b>3.6.1.</b> |
| |
| Feature vectorization |
| |
| </a> |
| |
| |
| |
| </li> |
| |
| <li class="chapter " data-level="3.6.2" data-path="../ft_engineering/quantify.html"> |
| |
| <a href="../ft_engineering/quantify.html"> |
| |
| |
| <b>3.6.2.</b> |
| |
| Quantify non-number features |
| |
| </a> |
| |
| |
| |
| </li> |
| |
| <li class="chapter " data-level="3.6.3" data-path="../ft_engineering/binarize.html"> |
| |
| <a href="../ft_engineering/binarize.html"> |
| |
| |
| <b>3.6.3.</b> |
| |
| Binarize label |
| |
| </a> |
| |
| |
| |
| </li> |
| |
| <li class="chapter " data-level="3.6.4" data-path="../ft_engineering/onehot.html"> |
| |
| <a href="../ft_engineering/onehot.html"> |
| |
| |
| <b>3.6.4.</b> |
| |
| One-hot encoding |
| |
| </a> |
| |
| |
| |
| </li> |
| |
| |
| </ul> |
| |
| </li> |
| |
| <li class="chapter " data-level="3.7" data-path="../ft_engineering/term_vector.html"> |
| |
| <a href="../ft_engineering/term_vector.html"> |
| |
| |
| <b>3.7.</b> |
| |
| Term Vector Model |
| |
| </a> |
| |
| |
| |
| <ul class="articles"> |
| |
| |
| <li class="chapter " data-level="3.7.1" data-path="../ft_engineering/tfidf.html"> |
| |
| <a href="../ft_engineering/tfidf.html"> |
| |
| |
| <b>3.7.1.</b> |
| |
| TF-IDF Term Weighting |
| |
| </a> |
| |
| |
| |
| </li> |
| |
| <li class="chapter " data-level="3.7.2" data-path="../ft_engineering/bm25.html"> |
| |
| <a href="../ft_engineering/bm25.html"> |
| |
| |
| <b>3.7.2.</b> |
| |
| Okapi BM25 Term Weighting |
| |
| </a> |
| |
| |
| |
| </li> |
| |
| |
| </ul> |
| |
| </li> |
| |
| |
| |
| |
| <li class="header">Part IV - Evaluation</li> |
| |
| |
| |
| <li class="chapter " data-level="4.1" data-path="../eval/binary_classification_measures.html"> |
| |
| <a href="../eval/binary_classification_measures.html"> |
| |
| |
| <b>4.1.</b> |
| |
| Binary Classification Metrics |
| |
| </a> |
| |
| |
| |
| <ul class="articles"> |
| |
| |
| <li class="chapter " data-level="4.1.1" data-path="../eval/auc.html"> |
| |
| <a href="../eval/auc.html"> |
| |
| |
| <b>4.1.1.</b> |
| |
| Area under the ROC curve |
| |
| </a> |
| |
| |
| |
| </li> |
| |
| |
| </ul> |
| |
| </li> |
| |
| <li class="chapter " data-level="4.2" data-path="../eval/multilabel_classification_measures.html"> |
| |
| <a href="../eval/multilabel_classification_measures.html"> |
| |
| |
| <b>4.2.</b> |
| |
| Multi-label Classification Metrics |
| |
| </a> |
| |
| |
| |
| </li> |
| |
| <li class="chapter " data-level="4.3" data-path="../eval/regression.html"> |
| |
| <a href="../eval/regression.html"> |
| |
| |
| <b>4.3.</b> |
| |
| Regression Metrics |
| |
| </a> |
| |
| |
| |
| </li> |
| |
| <li class="chapter " data-level="4.4" data-path="../eval/rank.html"> |
| |
| <a href="../eval/rank.html"> |
| |
| |
| <b>4.4.</b> |
| |
| Ranking Measures |
| |
| </a> |
| |
| |
| |
| </li> |
| |
| <li class="chapter " data-level="4.5" data-path="../eval/datagen.html"> |
| |
| <a href="../eval/datagen.html"> |
| |
| |
| <b>4.5.</b> |
| |
| Data Generation |
| |
| </a> |
| |
| |
| |
| <ul class="articles"> |
| |
| |
| <li class="chapter " data-level="4.5.1" data-path="../eval/lr_datagen.html"> |
| |
| <a href="../eval/lr_datagen.html"> |
| |
| |
| <b>4.5.1.</b> |
| |
| Logistic Regression data generation |
| |
| </a> |
| |
| |
| |
| </li> |
| |
| |
| </ul> |
| |
| </li> |
| |
| |
| |
| |
| <li class="header">Part V - Supervised Learning</li> |
| |
| |
| |
| <li class="chapter " data-level="5.1" data-path="../supervised_learning/prediction.html"> |
| |
| <a href="../supervised_learning/prediction.html"> |
| |
| |
| <b>5.1.</b> |
| |
| How Prediction Works |
| |
| </a> |
| |
| |
| |
| </li> |
| |
| <li class="chapter " data-level="5.2" data-path="../supervised_learning/tutorial.html"> |
| |
| <a href="../supervised_learning/tutorial.html"> |
| |
| |
| <b>5.2.</b> |
| |
| Step-by-Step Tutorial on Supervised Learning |
| |
| </a> |
| |
| |
| |
| </li> |
| |
| |
| |
| |
| <li class="header">Part VI - Binary Classification</li> |
| |
| |
| |
| <li class="chapter " data-level="6.1" data-path="general.html"> |
| |
| <a href="general.html"> |
| |
| |
| <b>6.1.</b> |
| |
| Binary Classification |
| |
| </a> |
| |
| |
| |
| </li> |
| |
| <li class="chapter " data-level="6.2" data-path="a9a.html"> |
| |
| <a href="a9a.html"> |
| |
| |
| <b>6.2.</b> |
| |
| a9a Tutorial |
| |
| </a> |
| |
| |
| |
| <ul class="articles"> |
| |
| |
| <li class="chapter " data-level="6.2.1" data-path="a9a_dataset.html"> |
| |
| <a href="a9a_dataset.html"> |
| |
| |
| <b>6.2.1.</b> |
| |
| Data Preparation |
| |
| </a> |
| |
| |
| |
| </li> |
| |
| <li class="chapter " data-level="6.2.2" data-path="a9a_generic.html"> |
| |
| <a href="a9a_generic.html"> |
| |
| |
| <b>6.2.2.</b> |
| |
| General Binary Classifier |
| |
| </a> |
| |
| |
| |
| </li> |
| |
| <li class="chapter " data-level="6.2.3" data-path="a9a_lr.html"> |
| |
| <a href="a9a_lr.html"> |
| |
| |
| <b>6.2.3.</b> |
| |
| Logistic Regression |
| |
| </a> |
| |
| |
| |
| </li> |
| |
| <li class="chapter " data-level="6.2.4" data-path="a9a_minibatch.html"> |
| |
| <a href="a9a_minibatch.html"> |
| |
| |
| <b>6.2.4.</b> |
| |
| Mini-batch Gradient Descent |
| |
| </a> |
| |
| |
| |
| </li> |
| |
| |
| </ul> |
| |
| </li> |
| |
| <li class="chapter " data-level="6.3" data-path="news20.html"> |
| |
| <a href="news20.html"> |
| |
| |
| <b>6.3.</b> |
| |
| News20 Tutorial |
| |
| </a> |
| |
| |
| |
| <ul class="articles"> |
| |
| |
| <li class="chapter " data-level="6.3.1" data-path="news20_dataset.html"> |
| |
| <a href="news20_dataset.html"> |
| |
| |
| <b>6.3.1.</b> |
| |
| Data Preparation |
| |
| </a> |
| |
| |
| |
| </li> |
| |
| <li class="chapter " data-level="6.3.2" data-path="news20_pa.html"> |
| |
| <a href="news20_pa.html"> |
| |
| |
| <b>6.3.2.</b> |
| |
| Perceptron, Passive Aggressive |
| |
| </a> |
| |
| |
| |
| </li> |
| |
| <li class="chapter " data-level="6.3.3" data-path="news20_scw.html"> |
| |
| <a href="news20_scw.html"> |
| |
| |
| <b>6.3.3.</b> |
| |
| CW, AROW, SCW |
| |
| </a> |
| |
| |
| |
| </li> |
| |
| <li class="chapter " data-level="6.3.4" data-path="news20_generic.html"> |
| |
| <a href="news20_generic.html"> |
| |
| |
| <b>6.3.4.</b> |
| |
| General Binary Classifier |
| |
| </a> |
| |
| |
| |
| </li> |
| |
| <li class="chapter " data-level="6.3.5" data-path="news20_generic_bagging.html"> |
| |
| <a href="news20_generic_bagging.html"> |
| |
| |
| <b>6.3.5.</b> |
| |
| Baggnig classiers |
| |
| </a> |
| |
| |
| |
| </li> |
| |
| <li class="chapter " data-level="6.3.6" data-path="news20_adagrad.html"> |
| |
| <a href="news20_adagrad.html"> |
| |
| |
| <b>6.3.6.</b> |
| |
| AdaGradRDA, AdaGrad, AdaDelta |
| |
| </a> |
| |
| |
| |
| </li> |
| |
| <li class="chapter " data-level="6.3.7" data-path="news20_rf.html"> |
| |
| <a href="news20_rf.html"> |
| |
| |
| <b>6.3.7.</b> |
| |
| Random Forest |
| |
| </a> |
| |
| |
| |
| </li> |
| |
| <li class="chapter active" data-level="6.3.8" data-path="news20b_xgboost.html"> |
| |
| <a href="news20b_xgboost.html"> |
| |
| |
| <b>6.3.8.</b> |
| |
| XGBoost |
| |
| </a> |
| |
| |
| |
| </li> |
| |
| |
| </ul> |
| |
| </li> |
| |
| <li class="chapter " data-level="6.4" data-path="kdd2010a.html"> |
| |
| <a href="kdd2010a.html"> |
| |
| |
| <b>6.4.</b> |
| |
| KDD2010a Tutorial |
| |
| </a> |
| |
| |
| |
| <ul class="articles"> |
| |
| |
| <li class="chapter " data-level="6.4.1" data-path="kdd2010a_dataset.html"> |
| |
| <a href="kdd2010a_dataset.html"> |
| |
| |
| <b>6.4.1.</b> |
| |
| Data Preparation |
| |
| </a> |
| |
| |
| |
| </li> |
| |
| <li class="chapter " data-level="6.4.2" data-path="kdd2010a_scw.html"> |
| |
| <a href="kdd2010a_scw.html"> |
| |
| |
| <b>6.4.2.</b> |
| |
| PA, CW, AROW, SCW |
| |
| </a> |
| |
| |
| |
| </li> |
| |
| |
| </ul> |
| |
| </li> |
| |
| <li class="chapter " data-level="6.5" data-path="kdd2010b.html"> |
| |
| <a href="kdd2010b.html"> |
| |
| |
| <b>6.5.</b> |
| |
| KDD2010b Tutorial |
| |
| </a> |
| |
| |
| |
| <ul class="articles"> |
| |
| |
| <li class="chapter " data-level="6.5.1" data-path="kdd2010b_dataset.html"> |
| |
| <a href="kdd2010b_dataset.html"> |
| |
| |
| <b>6.5.1.</b> |
| |
| Data Preparation |
| |
| </a> |
| |
| |
| |
| </li> |
| |
| <li class="chapter " data-level="6.5.2" data-path="kdd2010b_arow.html"> |
| |
| <a href="kdd2010b_arow.html"> |
| |
| |
| <b>6.5.2.</b> |
| |
| AROW |
| |
| </a> |
| |
| |
| |
| </li> |
| |
| |
| </ul> |
| |
| </li> |
| |
| <li class="chapter " data-level="6.6" data-path="webspam.html"> |
| |
| <a href="webspam.html"> |
| |
| |
| <b>6.6.</b> |
| |
| Webspam Tutorial |
| |
| </a> |
| |
| |
| |
| <ul class="articles"> |
| |
| |
| <li class="chapter " data-level="6.6.1" data-path="webspam_dataset.html"> |
| |
| <a href="webspam_dataset.html"> |
| |
| |
| <b>6.6.1.</b> |
| |
| Data Pareparation |
| |
| </a> |
| |
| |
| |
| </li> |
| |
| <li class="chapter " data-level="6.6.2" data-path="webspam_scw.html"> |
| |
| <a href="webspam_scw.html"> |
| |
| |
| <b>6.6.2.</b> |
| |
| PA1, AROW, SCW |
| |
| </a> |
| |
| |
| |
| </li> |
| |
| |
| </ul> |
| |
| </li> |
| |
| <li class="chapter " data-level="6.7" data-path="titanic_rf.html"> |
| |
| <a href="titanic_rf.html"> |
| |
| |
| <b>6.7.</b> |
| |
| Kaggle Titanic Tutorial |
| |
| </a> |
| |
| |
| |
| </li> |
| |
| <li class="chapter " data-level="6.8" data-path="criteo.html"> |
| |
| <a href="criteo.html"> |
| |
| |
| <b>6.8.</b> |
| |
| Criteo Tutorial |
| |
| </a> |
| |
| |
| |
| <ul class="articles"> |
| |
| |
| <li class="chapter " data-level="6.8.1" data-path="criteo_dataset.html"> |
| |
| <a href="criteo_dataset.html"> |
| |
| |
| <b>6.8.1.</b> |
| |
| Data Preparation |
| |
| </a> |
| |
| |
| |
| </li> |
| |
| <li class="chapter " data-level="6.8.2" data-path="criteo_ffm.html"> |
| |
| <a href="criteo_ffm.html"> |
| |
| |
| <b>6.8.2.</b> |
| |
| Field-Aware Factorization Machines |
| |
| </a> |
| |
| |
| |
| </li> |
| |
| |
| </ul> |
| |
| </li> |
| |
| |
| |
| |
| <li class="header">Part VII - Multiclass Classification</li> |
| |
| |
| |
| <li class="chapter " data-level="7.1" data-path="../multiclass/news20.html"> |
| |
| <a href="../multiclass/news20.html"> |
| |
| |
| <b>7.1.</b> |
| |
| News20 Multiclass Tutorial |
| |
| </a> |
| |
| |
| |
| <ul class="articles"> |
| |
| |
| <li class="chapter " data-level="7.1.1" data-path="../multiclass/news20_dataset.html"> |
| |
| <a href="../multiclass/news20_dataset.html"> |
| |
| |
| <b>7.1.1.</b> |
| |
| Data Preparation |
| |
| </a> |
| |
| |
| |
| </li> |
| |
| <li class="chapter " data-level="7.1.2" data-path="../multiclass/news20_one-vs-the-rest_dataset.html"> |
| |
| <a href="../multiclass/news20_one-vs-the-rest_dataset.html"> |
| |
| |
| <b>7.1.2.</b> |
| |
| Data Preparation for one-vs-the-rest classifiers |
| |
| </a> |
| |
| |
| |
| </li> |
| |
| <li class="chapter " data-level="7.1.3" data-path="../multiclass/news20_pa.html"> |
| |
| <a href="../multiclass/news20_pa.html"> |
| |
| |
| <b>7.1.3.</b> |
| |
| PA |
| |
| </a> |
| |
| |
| |
| </li> |
| |
| <li class="chapter " data-level="7.1.4" data-path="../multiclass/news20_scw.html"> |
| |
| <a href="../multiclass/news20_scw.html"> |
| |
| |
| <b>7.1.4.</b> |
| |
| CW, AROW, SCW |
| |
| </a> |
| |
| |
| |
| </li> |
| |
| <li class="chapter " data-level="7.1.5" data-path="../multiclass/news20_xgboost.html"> |
| |
| <a href="../multiclass/news20_xgboost.html"> |
| |
| |
| <b>7.1.5.</b> |
| |
| XGBoost |
| |
| </a> |
| |
| |
| |
| </li> |
| |
| <li class="chapter " data-level="7.1.6" data-path="../multiclass/news20_ensemble.html"> |
| |
| <a href="../multiclass/news20_ensemble.html"> |
| |
| |
| <b>7.1.6.</b> |
| |
| Ensemble learning |
| |
| </a> |
| |
| |
| |
| </li> |
| |
| <li class="chapter " data-level="7.1.7" data-path="../multiclass/news20_one-vs-the-rest.html"> |
| |
| <a href="../multiclass/news20_one-vs-the-rest.html"> |
| |
| |
| <b>7.1.7.</b> |
| |
| one-vs-the-rest Classifier |
| |
| </a> |
| |
| |
| |
| </li> |
| |
| |
| </ul> |
| |
| </li> |
| |
| <li class="chapter " data-level="7.2" data-path="../multiclass/iris.html"> |
| |
| <a href="../multiclass/iris.html"> |
| |
| |
| <b>7.2.</b> |
| |
| Iris Tutorial |
| |
| </a> |
| |
| |
| |
| <ul class="articles"> |
| |
| |
| <li class="chapter " data-level="7.2.1" data-path="../multiclass/iris_dataset.html"> |
| |
| <a href="../multiclass/iris_dataset.html"> |
| |
| |
| <b>7.2.1.</b> |
| |
| Data preparation |
| |
| </a> |
| |
| |
| |
| </li> |
| |
| <li class="chapter " data-level="7.2.2" data-path="../multiclass/iris_scw.html"> |
| |
| <a href="../multiclass/iris_scw.html"> |
| |
| |
| <b>7.2.2.</b> |
| |
| SCW |
| |
| </a> |
| |
| |
| |
| </li> |
| |
| <li class="chapter " data-level="7.2.3" data-path="../multiclass/iris_randomforest.html"> |
| |
| <a href="../multiclass/iris_randomforest.html"> |
| |
| |
| <b>7.2.3.</b> |
| |
| Random Forest |
| |
| </a> |
| |
| |
| |
| </li> |
| |
| <li class="chapter " data-level="7.2.4" data-path="../multiclass/iris_xgboost.html"> |
| |
| <a href="../multiclass/iris_xgboost.html"> |
| |
| |
| <b>7.2.4.</b> |
| |
| XGBoost |
| |
| </a> |
| |
| |
| |
| </li> |
| |
| |
| </ul> |
| |
| </li> |
| |
| |
| |
| |
| <li class="header">Part VIII - Regression</li> |
| |
| |
| |
| <li class="chapter " data-level="8.1" data-path="../regression/general.html"> |
| |
| <a href="../regression/general.html"> |
| |
| |
| <b>8.1.</b> |
| |
| Regression |
| |
| </a> |
| |
| |
| |
| </li> |
| |
| <li class="chapter " data-level="8.2" data-path="../regression/e2006.html"> |
| |
| <a href="../regression/e2006.html"> |
| |
| |
| <b>8.2.</b> |
| |
| E2006-tfidf Regression Tutorial |
| |
| </a> |
| |
| |
| |
| <ul class="articles"> |
| |
| |
| <li class="chapter " data-level="8.2.1" data-path="../regression/e2006_dataset.html"> |
| |
| <a href="../regression/e2006_dataset.html"> |
| |
| |
| <b>8.2.1.</b> |
| |
| Data Preparation |
| |
| </a> |
| |
| |
| |
| </li> |
| |
| <li class="chapter " data-level="8.2.2" data-path="../regression/e2006_generic.html"> |
| |
| <a href="../regression/e2006_generic.html"> |
| |
| |
| <b>8.2.2.</b> |
| |
| General Regessor |
| |
| </a> |
| |
| |
| |
| </li> |
| |
| <li class="chapter " data-level="8.2.3" data-path="../regression/e2006_arow.html"> |
| |
| <a href="../regression/e2006_arow.html"> |
| |
| |
| <b>8.2.3.</b> |
| |
| Passive Aggressive, AROW |
| |
| </a> |
| |
| |
| |
| </li> |
| |
| <li class="chapter " data-level="8.2.4" data-path="../regression/e2006_xgboost.html"> |
| |
| <a href="../regression/e2006_xgboost.html"> |
| |
| |
| <b>8.2.4.</b> |
| |
| XGBoost |
| |
| </a> |
| |
| |
| |
| </li> |
| |
| |
| </ul> |
| |
| </li> |
| |
| <li class="chapter " data-level="8.3" data-path="../regression/kddcup12tr2.html"> |
| |
| <a href="../regression/kddcup12tr2.html"> |
| |
| |
| <b>8.3.</b> |
| |
| KDDCup 2012 Track 2 CTR Prediction Tutorial |
| |
| </a> |
| |
| |
| |
| <ul class="articles"> |
| |
| |
| <li class="chapter " data-level="8.3.1" data-path="../regression/kddcup12tr2_dataset.html"> |
| |
| <a href="../regression/kddcup12tr2_dataset.html"> |
| |
| |
| <b>8.3.1.</b> |
| |
| Data Preparation |
| |
| </a> |
| |
| |
| |
| </li> |
| |
| <li class="chapter " data-level="8.3.2" data-path="../regression/kddcup12tr2_lr.html"> |
| |
| <a href="../regression/kddcup12tr2_lr.html"> |
| |
| |
| <b>8.3.2.</b> |
| |
| Logistic Regression, Passive Aggressive |
| |
| </a> |
| |
| |
| |
| </li> |
| |
| <li class="chapter " data-level="8.3.3" data-path="../regression/kddcup12tr2_lr_amplify.html"> |
| |
| <a href="../regression/kddcup12tr2_lr_amplify.html"> |
| |
| |
| <b>8.3.3.</b> |
| |
| Logistic Regression with amplifier |
| |
| </a> |
| |
| |
| |
| </li> |
| |
| <li class="chapter " data-level="8.3.4" data-path="../regression/kddcup12tr2_adagrad.html"> |
| |
| <a href="../regression/kddcup12tr2_adagrad.html"> |
| |
| |
| <b>8.3.4.</b> |
| |
| AdaGrad, AdaDelta |
| |
| </a> |
| |
| |
| |
| </li> |
| |
| |
| </ul> |
| |
| </li> |
| |
| |
| |
| |
| <li class="header">Part IX - Recommendation</li> |
| |
| |
| |
| <li class="chapter " data-level="9.1" data-path="../recommend/cf.html"> |
| |
| <a href="../recommend/cf.html"> |
| |
| |
| <b>9.1.</b> |
| |
| Collaborative Filtering |
| |
| </a> |
| |
| |
| |
| <ul class="articles"> |
| |
| |
| <li class="chapter " data-level="9.1.1" data-path="../recommend/item_based_cf.html"> |
| |
| <a href="../recommend/item_based_cf.html"> |
| |
| |
| <b>9.1.1.</b> |
| |
| Item-based Collaborative Filtering |
| |
| </a> |
| |
| |
| |
| </li> |
| |
| |
| </ul> |
| |
| </li> |
| |
| <li class="chapter " data-level="9.2" data-path="../recommend/news20.html"> |
| |
| <a href="../recommend/news20.html"> |
| |
| |
| <b>9.2.</b> |
| |
| News20 Related Article Recommendation Tutorial |
| |
| </a> |
| |
| |
| |
| <ul class="articles"> |
| |
| |
| <li class="chapter " data-level="9.2.1" data-path="../multiclass/news20_dataset.html"> |
| |
| <a href="../multiclass/news20_dataset.html"> |
| |
| |
| <b>9.2.1.</b> |
| |
| Data Preparation |
| |
| </a> |
| |
| |
| |
| </li> |
| |
| <li class="chapter " data-level="9.2.2" data-path="../recommend/news20_jaccard.html"> |
| |
| <a href="../recommend/news20_jaccard.html"> |
| |
| |
| <b>9.2.2.</b> |
| |
| LSH/MinHash and Jaccard Similarity |
| |
| </a> |
| |
| |
| |
| </li> |
| |
| <li class="chapter " data-level="9.2.3" data-path="../recommend/news20_knn.html"> |
| |
| <a href="../recommend/news20_knn.html"> |
| |
| |
| <b>9.2.3.</b> |
| |
| LSH/MinHash and Brute-force Search |
| |
| </a> |
| |
| |
| |
| </li> |
| |
| <li class="chapter " data-level="9.2.4" data-path="../recommend/news20_bbit_minhash.html"> |
| |
| <a href="../recommend/news20_bbit_minhash.html"> |
| |
| |
| <b>9.2.4.</b> |
| |
| kNN search using b-Bits MinHash |
| |
| </a> |
| |
| |
| |
| </li> |
| |
| |
| </ul> |
| |
| </li> |
| |
| <li class="chapter " data-level="9.3" data-path="../recommend/movielens.html"> |
| |
| <a href="../recommend/movielens.html"> |
| |
| |
| <b>9.3.</b> |
| |
| MovieLens Movie Recommendation Tutorial |
| |
| </a> |
| |
| |
| |
| <ul class="articles"> |
| |
| |
| <li class="chapter " data-level="9.3.1" data-path="../recommend/movielens_dataset.html"> |
| |
| <a href="../recommend/movielens_dataset.html"> |
| |
| |
| <b>9.3.1.</b> |
| |
| Data Preparation |
| |
| </a> |
| |
| |
| |
| </li> |
| |
| <li class="chapter " data-level="9.3.2" data-path="../recommend/movielens_cf.html"> |
| |
| <a href="../recommend/movielens_cf.html"> |
| |
| |
| <b>9.3.2.</b> |
| |
| Item-based Collaborative Filtering |
| |
| </a> |
| |
| |
| |
| </li> |
| |
| <li class="chapter " data-level="9.3.3" data-path="../recommend/movielens_mf.html"> |
| |
| <a href="../recommend/movielens_mf.html"> |
| |
| |
| <b>9.3.3.</b> |
| |
| Matrix Factorization |
| |
| </a> |
| |
| |
| |
| </li> |
| |
| <li class="chapter " data-level="9.3.4" data-path="../recommend/movielens_fm.html"> |
| |
| <a href="../recommend/movielens_fm.html"> |
| |
| |
| <b>9.3.4.</b> |
| |
| Factorization Machine |
| |
| </a> |
| |
| |
| |
| </li> |
| |
| <li class="chapter " data-level="9.3.5" data-path="../recommend/movielens_slim.html"> |
| |
| <a href="../recommend/movielens_slim.html"> |
| |
| |
| <b>9.3.5.</b> |
| |
| SLIM for fast top-k Recommendation |
| |
| </a> |
| |
| |
| |
| </li> |
| |
| <li class="chapter " data-level="9.3.6" data-path="../recommend/movielens_cv.html"> |
| |
| <a href="../recommend/movielens_cv.html"> |
| |
| |
| <b>9.3.6.</b> |
| |
| 10-fold Cross Validation (Matrix Factorization) |
| |
| </a> |
| |
| |
| |
| </li> |
| |
| |
| </ul> |
| |
| </li> |
| |
| |
| |
| |
| <li class="header">Part X - Anomaly Detection</li> |
| |
| |
| |
| <li class="chapter " data-level="10.1" data-path="../anomaly/lof.html"> |
| |
| <a href="../anomaly/lof.html"> |
| |
| |
| <b>10.1.</b> |
| |
| Outlier Detection using Local Outlier Factor (LOF) |
| |
| </a> |
| |
| |
| |
| </li> |
| |
| <li class="chapter " data-level="10.2" data-path="../anomaly/sst.html"> |
| |
| <a href="../anomaly/sst.html"> |
| |
| |
| <b>10.2.</b> |
| |
| Change-Point Detection using Singular Spectrum Transformation (SST) |
| |
| </a> |
| |
| |
| |
| </li> |
| |
| <li class="chapter " data-level="10.3" data-path="../anomaly/changefinder.html"> |
| |
| <a href="../anomaly/changefinder.html"> |
| |
| |
| <b>10.3.</b> |
| |
| ChangeFinder: Detecting Outlier and Change-Point Simultaneously |
| |
| </a> |
| |
| |
| |
| </li> |
| |
| |
| |
| |
| <li class="header">Part XI - Clustering</li> |
| |
| |
| |
| <li class="chapter " data-level="11.1" data-path="../clustering/lda.html"> |
| |
| <a href="../clustering/lda.html"> |
| |
| |
| <b>11.1.</b> |
| |
| Latent Dirichlet Allocation |
| |
| </a> |
| |
| |
| |
| </li> |
| |
| <li class="chapter " data-level="11.2" data-path="../clustering/plsa.html"> |
| |
| <a href="../clustering/plsa.html"> |
| |
| |
| <b>11.2.</b> |
| |
| Probabilistic Latent Semantic Analysis |
| |
| </a> |
| |
| |
| |
| </li> |
| |
| |
| |
| |
| <li class="header">Part XII - GeoSpatial Functions</li> |
| |
| |
| |
| <li class="chapter " data-level="12.1" data-path="../geospatial/latlon.html"> |
| |
| <a href="../geospatial/latlon.html"> |
| |
| |
| <b>12.1.</b> |
| |
| Lat/Lon functions |
| |
| </a> |
| |
| |
| |
| </li> |
| |
| |
| |
| |
| <li class="header">Part XIII - Hivemall on SparkSQL</li> |
| |
| |
| |
| <li class="chapter " data-level="13.1" data-path="../spark/getting_started/README.md"> |
| |
| <span> |
| |
| |
| <b>13.1.</b> |
| |
| Getting Started |
| |
| </a> |
| |
| |
| |
| <ul class="articles"> |
| |
| |
| <li class="chapter " data-level="13.1.1" data-path="../spark/getting_started/installation.html"> |
| |
| <a href="../spark/getting_started/installation.html"> |
| |
| |
| <b>13.1.1.</b> |
| |
| Installation |
| |
| </a> |
| |
| |
| |
| </li> |
| |
| |
| </ul> |
| |
| </li> |
| |
| <li class="chapter " data-level="13.2" data-path="../spark/binaryclass/"> |
| |
| <a href="../spark/binaryclass/"> |
| |
| |
| <b>13.2.</b> |
| |
| Binary Classification |
| |
| </a> |
| |
| |
| |
| <ul class="articles"> |
| |
| |
| <li class="chapter " data-level="13.2.1" data-path="../spark/binaryclass/a9a_sql.html"> |
| |
| <a href="../spark/binaryclass/a9a_sql.html"> |
| |
| |
| <b>13.2.1.</b> |
| |
| a9a Tutorial for SQL |
| |
| </a> |
| |
| |
| |
| </li> |
| |
| |
| </ul> |
| |
| </li> |
| |
| <li class="chapter " data-level="13.3" data-path="../spark/binaryclass/"> |
| |
| <a href="../spark/binaryclass/"> |
| |
| |
| <b>13.3.</b> |
| |
| Regression |
| |
| </a> |
| |
| |
| |
| <ul class="articles"> |
| |
| |
| <li class="chapter " data-level="13.3.1" data-path="../spark/regression/e2006_sql.html"> |
| |
| <a href="../spark/regression/e2006_sql.html"> |
| |
| |
| <b>13.3.1.</b> |
| |
| E2006-tfidf Regression Tutorial for SQL |
| |
| </a> |
| |
| |
| |
| </li> |
| |
| |
| </ul> |
| |
| </li> |
| |
| |
| |
| |
| <li class="header">Part XIV - Hivemall on Docker</li> |
| |
| |
| |
| <li class="chapter " data-level="14.1" data-path="../docker/getting_started.html"> |
| |
| <a href="../docker/getting_started.html"> |
| |
| |
| <b>14.1.</b> |
| |
| Getting Started |
| |
| </a> |
| |
| |
| |
| </li> |
| |
| |
| |
| |
| <li class="header">Part XIV - External References</li> |
| |
| |
| |
| <li class="chapter " data-level="15.1" > |
| |
| <a target="_blank" href="https://github.com/daijyc/hivemall/wiki/PigHome"> |
| |
| |
| <b>15.1.</b> |
| |
| Hivemall on Apache Pig |
| |
| </a> |
| |
| |
| |
| </li> |
| |
| |
| |
| |
| <li class="divider"></li> |
| |
| <li> |
| <a href="https://www.gitbook.com" target="blank" class="gitbook-link"> |
| Published with GitBook |
| </a> |
| </li> |
| </ul> |
| |
| |
| </nav> |
| |
| |
| </div> |
| |
| <div class="book-body"> |
| |
| <div class="body-inner"> |
| |
| |
| |
| <div class="book-header" role="navigation"> |
| |
| |
| <!-- Title --> |
| <h1> |
| <i class="fa fa-circle-o-notch fa-spin"></i> |
| <a href=".." >XGBoost</a> |
| </h1> |
| </div> |
| |
| |
| |
| |
| <div class="page-wrapper" tabindex="-1" role="main"> |
| <div class="page-inner"> |
| |
| <div id="book-search-results"> |
| <div class="search-noresults"> |
| |
| <section class="normal markdown-section"> |
| |
| <!-- |
| 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. |
| --> |
| <p>In this tutorial, we build a binary classification model using XGBoost.</p> |
| <!-- toc --><div id="toc" class="toc"> |
| |
| <ul> |
| <li><a href="#feature-vector-format-for-xgboost">Feature Vector format for XGBoost</a></li> |
| <li><a href="#label-format-in-binary-classification">Label format in Binary Classification</a></li> |
| <li><a href="#usage-and-hyperparameters">Usage and Hyperparameters</a></li> |
| <li><a href="#training">Training</a></li> |
| <li><a href="#prediction">prediction</a></li> |
| <li><a href="#evaluation">evaluation</a></li> |
| </ul> |
| |
| </div><!-- tocstop --> |
| <h2 id="feature-vector-format-for-xgboost">Feature Vector format for XGBoost</h2> |
| <p>For feature vector, <code>train_xgboost</code> takes a sparse vector format (<code>array<string></code>) or a dense vector format (<code>array<double></code>). |
| In the feature vector, each feature takes a LIBSVM format:</p> |
| <pre><code>feature ::= <index>:<weight> |
| |
| index ::= <Non-negative INT> (e.g., 0,1,2,...) |
| weight ::= <DOUBLE> |
| </code></pre><div class="panel panel-primary"><div class="panel-heading"><h3 class="panel-title" id="note"><i class="fa fa-edit"></i> Note</h3></div><div class="panel-body"><p>Unlike the original libsvm format, it's not needed to sort a feature vector by ansceding order of feature index.</p></div></div> |
| <p>Target label format of binary classification follows <a href="http://hivemall.apache.org/userguide/getting_started/input-format.html#label-format-in-binary-classification" target="_blank">this rule</a>. Please refer <a href="https://xgboost.readthedocs.io/en/latest/tutorials/input_format.html" target="_blank">xgboost document</a> as well.</p> |
| <h2 id="label-format-in-binary-classification">Label format in Binary Classification</h2> |
| <p>The label must be an INT typed column and the values are positive (+1) or negative (-1) as follows:</p> |
| <pre><code><label> ::= 1 | -1 |
| </code></pre><p>Alternatively, you can use the following format that represents 1 for a positive example and 0 for a negative example:</p> |
| <pre><code><label> ::= 0 | 1 |
| </code></pre><h2 id="usage-and-hyperparameters">Usage and Hyperparameters</h2> |
| <p>You can find hyperparameters and it's default setting by running the following query:</p> |
| <pre><code class="lang-sql">select train_xgboost(); |
| |
| usage: train_xgboost(array<string|double> features, int|double target [, |
| string options]) - Returns a relation consists of <string model_id, |
| array<string> pred_model> [-alpha <arg>] [-base_score <arg>] |
| [-booster <arg>] [-colsample_bylevel <arg>] [-colsample_bynode |
| <arg>] [-colsample_bytree <arg>] [-disable_default_eval_metric |
| <arg>] [-eta <arg>] [-eval_metric <arg>] [-feature_selector <arg>] |
| [-gamma <arg>] [-grow_policy <arg>] [-lambda <arg>] [-lambda_bias |
| <arg>] [-max_bin <arg>] [-max_delta_step <arg>] [-max_depth <arg>] |
| [-max_leaves <arg>] [-maximize_evaluation_metrics <arg>] |
| [-min_child_weight <arg>] [-normalize_type <arg>] [-num_class |
| <arg>] [-num_early_stopping_rounds <arg>] [-num_feature <arg>] |
| [-num_parallel_tree <arg>] [-num_pbuffer <arg>] [-num_round <arg>] |
| [-objective <arg>] [-one_drop <arg>] [-process_type <arg>] |
| [-rate_drop <arg>] [-refresh_leaf <arg>] [-sample_type <arg>] |
| [-scale_pos_weight <arg>] [-seed <arg>] [-silent <arg>] |
| [-sketch_eps <arg>] [-skip_drop <arg>] [-subsample <arg>] [-top_k |
| <arg>] [-tree_method <arg>] [-tweedie_variance_power <arg>] |
| [-updater <arg>] [-validation_ratio <arg>] [-verbosity <arg>] |
| -alpha,--reg_alpha <arg> L1 regularization term on weights. |
| Increasing this value will make |
| model more conservative. [default: |
| 0.0] |
| -base_score <arg> Initial prediction score of all |
| instances, global bias [default: |
| 0.5] |
| -booster <arg> Set a booster to use, gbtree or |
| gblinear or dart. [default: gbree] |
| -colsample_bylevel <arg> Subsample ratio of columns for each |
| level [default: 1.0] |
| -colsample_bynode <arg> Subsample ratio of columns for each |
| node [default: 1.0] |
| -colsample_bytree <arg> Subsample ratio of columns when |
| constructing each tree [default: |
| 1.0] |
| -disable_default_eval_metric <arg> NFlag to disable default metric. Set |
| to >0 to disable. [default: 0] |
| -eta,--learning_rate <arg> Step size shrinkage used in update |
| to prevents overfitting [default: |
| 0.3] |
| -eval_metric <arg> Evaluation metrics for validation |
| data. A default metric is assigned |
| according to the objective: |
| - rmse: for regression |
| - error: for classification |
| - map: for ranking |
| For a list of valid inputs, see |
| XGBoost Parameters. |
| -feature_selector <arg> Feature selection and ordering |
| method. [Choices: cyclic (default), |
| shuffle, random, greedy, thrifty] |
| -gamma,--min_split_loss <arg> Minimum loss reduction required to |
| make a further partition on a leaf |
| node of the tree. [default: 0.0] |
| -grow_policy <arg> Controls a way new nodes are added |
| to the tree. Currently supported |
| only if tree_method is set to hist. |
| [default: depthwise, Choices: |
| depthwise, lossguide] |
| -lambda,--reg_lambda <arg> L2 regularization term on weights. |
| Increasing this value will make |
| model more conservative. [default: |
| 1.0 for gbtree, 0.0 for gblinear] |
| -lambda_bias <arg> L2 regularization term on bias |
| [default: 0.0] |
| -max_bin <arg> Maximum number of discrete bins to |
| bucket continuous features. Only |
| used if tree_method is set to hist. |
| [default: 256] |
| -max_delta_step <arg> Maximum delta step we allow each |
| tree's weight estimation to be |
| [default: 0] |
| -max_depth <arg> Max depth of decision tree [default: |
| 6] |
| -max_leaves <arg> Maximum number of nodes to be added. |
| Only relevant when |
| grow_policy=lossguide is set. |
| [default: 0] |
| -maximize_evaluation_metrics <arg> Maximize evaluation metrics |
| [default: false] |
| -min_child_weight <arg> Minimum sum of instance weight |
| (hessian) needed in a child |
| [default: 1.0] |
| -normalize_type <arg> Type of normalization algorithm. |
| [Choices: tree (default), forest] |
| -num_class <arg> Number of classes to classify |
| -num_early_stopping_rounds <arg> Minimum rounds required for early |
| stopping [default: 0] |
| -num_feature <arg> Feature dimension used in boosting |
| [default: set automatically by |
| xgboost] |
| -num_parallel_tree <arg> Number of parallel trees constructed |
| during each iteration. This option |
| is used to support boosted random |
| forest. [default: 1] |
| -num_pbuffer <arg> Size of prediction buffer [default: |
| set automatically by xgboost] |
| -num_round,--iters <arg> Number of boosting iterations |
| [default: 10] |
| -objective <arg> Specifies the learning task and the |
| corresponding learning objective. |
| Examples: reg:linear, reg:logistic, |
| multi:softmax. For a full list of |
| valid inputs, refer to XGBoost |
| Parameters. [default: reg:linear] |
| -one_drop <arg> When this flag is enabled, at least |
| one tree is always dropped during |
| the dropout. 0 or 1. [default: 0] |
| -process_type <arg> A type of boosting process to run. |
| [Choices: default, update] |
| -rate_drop <arg> Dropout rate in range [0.0, 1.0]. |
| [default: 0.0] |
| -refresh_leaf <arg> This is a parameter of the refresh |
| updater plugin. When this flag is 1, |
| tree leafs as well as tree nodes’ |
| stats are updated. When it is 0, |
| only node stats are updated. |
| [default: 1] |
| -sample_type <arg> Type of sampling algorithm. |
| [Choices: uniform (default), |
| weighted] |
| -scale_pos_weight <arg> ontrol the balance of positive and |
| negative weights, useful for |
| unbalanced classes. A typical value |
| to consider: sum(negative instances) |
| / sum(positive instances) [default: |
| 1.0] |
| -seed <arg> Random number seed. [default: 43] |
| -silent <arg> Deprecated. Please use verbosity |
| instead. 0 means printing running |
| messages, 1 means silent mode |
| [default: 1] |
| -sketch_eps <arg> This roughly translates into O(1 / |
| sketch_eps) number of bins. |
| Compared to directly select number |
| of bins, this comes with theoretical |
| guarantee with sketch accuracy. |
| Only used for tree_method=approx. |
| Usually user does not have to tune |
| this. [default: 0.03] |
| -skip_drop <arg> Probability of skipping the dropout |
| procedure during a boosting |
| iteration in range [0.0, 1.0]. |
| [default: 0.0] |
| -subsample <arg> Subsample ratio of the training |
| instance in range (0.0,1.0] |
| [default: 1.0] |
| -top_k <arg> The number of top features to select |
| in greedy and thrifty feature |
| selector. The value of 0 means using |
| all the features. [default: 0] |
| -tree_method <arg> The tree construction algorithm used |
| in XGBoost. [default: auto, Choices: |
| auto, exact, approx, hist] |
| -tweedie_variance_power <arg> Parameter that controls the variance |
| of the Tweedie distribution in range |
| [1.0, 2.0]. [default: 1.5] |
| -updater <arg> A comma-separated string that |
| defines the sequence of tree |
| updaters to run. For a full list of |
| valid inputs, please refer to |
| XGBoost Parameters. [default: |
| 'grow_colmaker,prune' for gbtree, |
| 'shotgun' for gblinear] |
| -validation_ratio <arg> Validation ratio in range [0.0,1.0] |
| [default: 0.2] |
| -verbosity <arg> Verbosity of printing messages. |
| Choices: 0 (silent), 1 (warning), 2 |
| (info), 3 (debug). [default: 0] |
| </code></pre> |
| <p>Objective function <code>-objective</code> SHOULD be specified though <code>-objective reg:linear</code> is used for Objective function by the default. |
| For the full list of objective functions, please refer <a href="https://xgboost.readthedocs.io/en/stable/parameter.html#learning-task-parameters" target="_blank">this xgboost v0.90 documentation</a>.</p> |
| <p>The following objectives would widely be used for regression, binary classication, and multiclass classication, respectively.</p> |
| <ul> |
| <li><code>reg:squarederror</code> regression with squared loss.</li> |
| <li><code>binary:logistic</code> logistic regression for binary classification, output probability.</li> |
| <li><code>binary:hinge</code> hinge loss for binary classification. This makes predictions of 0 or 1, rather than producing probabilities.</li> |
| <li><code>multi:softmax</code> set XGBoost to do multiclass classification using the softmax objective, you also need to set <code>num_class</code> (number of classes).</li> |
| <li><code>multi:softprob</code> same as softmax, but output a vector of <code>ndata * nclass</code>, which can be further reshaped to <code>ndata * nclass</code> matrix. The result contains predicted probability of each data point belonging to each class.</li> |
| </ul> |
| <p>Other hyperparameters better to be tuned are:</p> |
| <ul> |
| <li><code>-booster gbree</code> Which booster to use. The default gbtree (Gradient Boosting Trees) would be fine for most cases. Can be <code>gbtree</code>, <code>gblinear</code> or <code>dart</code>; gbtree and dart use tree based models while gblinear uses linear functions.</li> |
| <li><code>-eta 0.1</code> The learning rate, 0.3 by the default. 0.05, 0.1, 0.3 are worth trying.</li> |
| <li><code>-max_depth 6</code> The maximum depth of the tree. The default value 6 would be fine for most case. Recommended value range is 5-10.</li> |
| <li><code>-num_class 3</code> The number of classes MUST be specified for multiclass classification (i.e., <code>-objective multi:softmax</code> or <code>-objective multi:softprob</code>)</li> |
| <li><code>-num_round 10</code> The number of rounds for boosting. 10 or more would be preferred.</li> |
| <li><code>-num_early_stopping_rounds 3</code> The number of rounds required for early stopping. Without specifying <code>-num_early_stopping_rounds</code>, no early stopping is NOT carried. When <code>-num_round=100</code> and <code>-num_early_stopping_rounds=5</code>, traning could be early stopped at 15th iteration if there is no evaluation result greater than the 10th iteration's (best one). Early stopping 3 or so would be preferred. </li> |
| <li><code>-validation_ratio 0.2</code> The ratio data used for validation (early stopping). 0.2 would be enough for most cases. Note that 80% data is used for training when <code>validation_ratio 0.2</code> is set.</li> |
| </ul> |
| <p>You can find the underlying XGBoost version by:</p> |
| <pre><code class="lang-sql">select xgboost_version(); |
| > 0.90 |
| </code></pre> |
| <h2 id="training">Training</h2> |
| <p><code>train_xgboost</code> UDTF is used for training. </p> |
| <p>The function signature is <code>train_xgboost(array<string|double> features, double target [,string options])</code> and it returns a prediction model as a relation consist of <code><string model_id, array<string> pred_model></code>.</p> |
| <pre><code class="lang-sql"><span class="hljs-comment">-- explicitly use 3 reducers</span> |
| <span class="hljs-comment">-- set mapred.reduce.tasks=3;</span> |
| |
| <span class="hljs-keyword">drop</span> <span class="hljs-keyword">table</span> xgb_lr_model; |
| <span class="hljs-keyword">create</span> <span class="hljs-keyword">table</span> xgb_lr_model <span class="hljs-keyword">as</span> |
| <span class="hljs-keyword">select</span> |
| train_xgboost(features, label, <span class="hljs-string">'-objective binary:logistic -num_round 10 -num_early_stopping_rounds 3'</span>) |
| <span class="hljs-keyword">as</span> (model_id, <span class="hljs-keyword">model</span>) |
| <span class="hljs-keyword">from</span> ( |
| <span class="hljs-keyword">select</span> features, label |
| <span class="hljs-keyword">from</span> news20b_train |
| cluster <span class="hljs-keyword">by</span> <span class="hljs-keyword">rand</span>(<span class="hljs-number">43</span>) <span class="hljs-comment">-- shuffle data to reducers</span> |
| ) shuffled; |
| |
| <span class="hljs-keyword">drop</span> <span class="hljs-keyword">table</span> xgb_hinge_model; |
| <span class="hljs-keyword">create</span> <span class="hljs-keyword">table</span> xgb_hinge_model <span class="hljs-keyword">as</span> |
| <span class="hljs-keyword">select</span> |
| train_xgboost(features, label, <span class="hljs-string">'-objective binary:hinge -num_round 10 -num_early_stopping_rounds 3'</span>) |
| <span class="hljs-keyword">as</span> (model_id, <span class="hljs-keyword">model</span>) |
| <span class="hljs-keyword">from</span> ( |
| <span class="hljs-keyword">select</span> features, label |
| <span class="hljs-keyword">from</span> news20b_train |
| cluster <span class="hljs-keyword">by</span> <span class="hljs-keyword">rand</span>(<span class="hljs-number">43</span>) <span class="hljs-comment">-- shuffle data to reducers</span> |
| ) shuffled; |
| </code></pre> |
| <div class="panel panel-warning"><div class="panel-heading"><h3 class="panel-title" id="caution"><i class="fa fa-exclamation-triangle"></i> Caution</h3></div><div class="panel-body"><p><code>cluster by rand()</code> is NOT required when training data is small and a single task is launched for XGBoost training. |
| <code>cluster by rand()</code> shuffles data at random and divided it for multiple XGBoost instances.</p></div></div> |
| <h2 id="prediction">prediction</h2> |
| <pre><code class="lang-sql"><span class="hljs-keyword">drop</span> <span class="hljs-keyword">table</span> xgb_lr_predicted; |
| <span class="hljs-keyword">create</span> <span class="hljs-keyword">table</span> xgb_lr_predicted |
| <span class="hljs-keyword">as</span> |
| <span class="hljs-keyword">select</span> |
| <span class="hljs-keyword">rowid</span>, |
| array_avg(predicted) <span class="hljs-keyword">as</span> predicted, |
| <span class="hljs-keyword">avg</span>(predicted[<span class="hljs-number">0</span>]) <span class="hljs-keyword">as</span> prob |
| <span class="hljs-keyword">from</span> ( |
| <span class="hljs-keyword">select</span> |
| <span class="hljs-comment">-- fast predictition by xgboost-predictor-java (https://github.com/komiya-atsushi/xgboost-predictor-java/)</span> |
| xgboost_predict(<span class="hljs-keyword">rowid</span>, features, model_id, <span class="hljs-keyword">model</span>) <span class="hljs-keyword">as</span> (<span class="hljs-keyword">rowid</span>, predicted) |
| <span class="hljs-comment">-- predict by xgboost4j (https://xgboost.readthedocs.io/en/stable/jvm/)</span> |
| <span class="hljs-comment">-- xgboost_batch_predict(rowid, features, model_id, model) as (rowid, predicted)</span> |
| <span class="hljs-keyword">from</span> |
| <span class="hljs-comment">-- for each model l </span> |
| <span class="hljs-comment">-- for each test r</span> |
| <span class="hljs-comment">-- predict</span> |
| xgb_lr_model l |
| <span class="hljs-keyword">LEFT</span> <span class="hljs-keyword">OUTER</span> <span class="hljs-keyword">JOIN</span> news20b_test r |
| ) t |
| <span class="hljs-keyword">group</span> <span class="hljs-keyword">by</span> <span class="hljs-keyword">rowid</span>; |
| |
| <span class="hljs-keyword">drop</span> <span class="hljs-keyword">table</span> xgb_hinge_predicted; |
| <span class="hljs-keyword">create</span> <span class="hljs-keyword">table</span> xgb_hinge_predicted |
| <span class="hljs-keyword">as</span> |
| <span class="hljs-keyword">select</span> |
| <span class="hljs-keyword">rowid</span>, |
| <span class="hljs-comment">-- voting</span> |
| <span class="hljs-comment">-- if(sum(if(predicted[0]=1,1,0)) > sum(if(predicted[0]=0,1,0)),1,-1) as predicted</span> |
| majority_vote(<span class="hljs-keyword">if</span>(predicted[<span class="hljs-number">0</span>]=<span class="hljs-number">1</span>, <span class="hljs-number">1</span>, <span class="hljs-number">-1</span>)) <span class="hljs-keyword">as</span> predicted |
| <span class="hljs-keyword">from</span> ( |
| <span class="hljs-keyword">select</span> |
| <span class="hljs-comment">-- binary:hinge is not supported in xgboost_predict</span> |
| <span class="hljs-comment">-- binary:hinge returns [1.0] or [0.0] for predicted</span> |
| xgboost_batch_predict(<span class="hljs-keyword">rowid</span>, features, model_id, <span class="hljs-keyword">model</span>) |
| <span class="hljs-keyword">as</span> (<span class="hljs-keyword">rowid</span>, predicted) |
| <span class="hljs-keyword">from</span> |
| <span class="hljs-comment">-- for each model l </span> |
| <span class="hljs-comment">-- for each test r</span> |
| <span class="hljs-comment">-- predict</span> |
| xgb_hinge_model l |
| <span class="hljs-keyword">LEFT</span> <span class="hljs-keyword">OUTER</span> <span class="hljs-keyword">JOIN</span> news20b_test r |
| ) t |
| <span class="hljs-keyword">group</span> <span class="hljs-keyword">by</span> |
| <span class="hljs-keyword">rowid</span> |
| </code></pre> |
| <p>You can find the function signature of <code>xgboost_predict</code> by</p> |
| <pre><code class="lang-sql">select xgboost_predict(); |
| |
| usage: xgboost_predict(PRIMITIVE rowid, array<string|double> features, |
| string model_id, array<string> pred_model [, string options]) - |
| Returns a prediction result as (string rowid, array<double> |
| predicted) |
| |
| select xgboost_batch_predict(); |
| |
| usage: xgboost_batch_predict(PRIMITIVE rowid, array<string|double> |
| features, string model_id, array<string> pred_model [, string |
| options]) - Returns a prediction result as (string rowid, |
| array<double> predicted) [-batch_size <arg>] |
| -batch_size <arg> Number of rows to predict together [default: 128] |
| </code></pre> |
| <div class="panel panel-warning"><div class="panel-heading"><h3 class="panel-title" id="caution"><i class="fa fa-exclamation-triangle"></i> Caution</h3></div><div class="panel-body"><p><code>xgboost_predict</code> outputs probability for <code>-objective binary:logistic</code> while 0/1 is resulted for <code>-objective binary:hinge</code>.</p><p><code>xgboost_predict</code> only support the following models and objectives because it uses <a href="https://github.com/komiya-atsushi/xgboost-predictor-java" target="_blank">xgboost-predictor-java</a>: |
| Models: {gblinear, gbtree, dart} |
| Objective functions: {binary:logistic, binary:logitraw, multi:softmax, multi:softprob, reg:linear, reg:squarederror, rank:pairwise}</p><p>For other models and objectives, please use <code>xgboost_batch_predict</code> that uses <a href="https://xgboost.readthedocs.io/en/stable/jvm/" target="_blank">xgboost4j</a> insead.</p></div></div> |
| <h2 id="evaluation">evaluation</h2> |
| <pre><code class="lang-sql">WITH submit as ( |
| <span class="hljs-keyword">select</span> |
| t.label <span class="hljs-keyword">as</span> actual, |
| <span class="hljs-comment">-- probability thresholding by 0.5</span> |
| <span class="hljs-keyword">if</span>(p.prob > <span class="hljs-number">0.5</span>,<span class="hljs-number">1</span>,<span class="hljs-number">-1</span>) <span class="hljs-keyword">as</span> predicted |
| <span class="hljs-keyword">from</span> |
| news20b_test t |
| <span class="hljs-keyword">JOIN</span> xgb_lr_predicted p |
| <span class="hljs-keyword">on</span> (t.<span class="hljs-keyword">rowid</span> = p.<span class="hljs-keyword">rowid</span>) |
| ) |
| <span class="hljs-keyword">select</span> |
| <span class="hljs-keyword">sum</span>(<span class="hljs-keyword">if</span>(actual = predicted, <span class="hljs-number">1</span>, <span class="hljs-number">0</span>)) / <span class="hljs-keyword">count</span>(<span class="hljs-number">1</span>) <span class="hljs-keyword">as</span> accuracy |
| <span class="hljs-keyword">from</span> |
| submit; |
| </code></pre> |
| <blockquote> |
| <p>0.8372698158526821 (logistic loss)</p> |
| </blockquote> |
| <pre><code class="lang-sql">WITH submit as ( |
| <span class="hljs-keyword">select</span> |
| t.label <span class="hljs-keyword">as</span> actual, |
| p.predicted |
| <span class="hljs-keyword">from</span> |
| news20b_test t |
| <span class="hljs-keyword">JOIN</span> xgb_hinge_predicted p |
| <span class="hljs-keyword">on</span> (t.<span class="hljs-keyword">rowid</span> = p.<span class="hljs-keyword">rowid</span>) |
| ) |
| <span class="hljs-keyword">select</span> |
| <span class="hljs-keyword">sum</span>(<span class="hljs-keyword">if</span>(actual=predicted,<span class="hljs-number">1</span>,<span class="hljs-number">0</span>)) / <span class="hljs-keyword">count</span>(<span class="hljs-number">1</span>) <span class="hljs-keyword">as</span> accuracy |
| <span class="hljs-keyword">from</span> |
| submit; |
| </code></pre> |
| <blockquote> |
| <p>0.7752201761409128 (hinge loss)</p> |
| </blockquote> |
| <p><div id="page-footer" class="localized-footer"><hr><!-- |
| 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. |
| --> |
| <p><sub><font color="gray"> |
| Apache Hivemall is an effort undergoing incubation at The Apache Software Foundation (ASF), sponsored by the Apache Incubator. |
| </font></sub></p> |
| </div></p> |
| |
| |
| </section> |
| |
| </div> |
| <div class="search-results"> |
| <div class="has-results"> |
| |
| <h1 class="search-results-title"><span class='search-results-count'></span> results matching "<span class='search-query'></span>"</h1> |
| <ul class="search-results-list"></ul> |
| |
| </div> |
| <div class="no-results"> |
| |
| <h1 class="search-results-title">No results matching "<span class='search-query'></span>"</h1> |
| |
| </div> |
| </div> |
| </div> |
| |
| </div> |
| </div> |
| |
| </div> |
| |
| |
| |
| |
| </div> |
| |
| <script> |
| var gitbook = gitbook || []; |
| gitbook.push(function() { |
| gitbook.page.hasChanged({"page":{"title":"XGBoost","level":"6.3.8","depth":2,"next":{"title":"KDD2010a Tutorial","level":"6.4","depth":1,"path":"binaryclass/kdd2010a.md","ref":"binaryclass/kdd2010a.md","articles":[{"title":"Data Preparation","level":"6.4.1","depth":2,"path":"binaryclass/kdd2010a_dataset.md","ref":"binaryclass/kdd2010a_dataset.md","articles":[]},{"title":"PA, CW, AROW, SCW","level":"6.4.2","depth":2,"path":"binaryclass/kdd2010a_scw.md","ref":"binaryclass/kdd2010a_scw.md","articles":[]}]},"previous":{"title":"Random Forest","level":"6.3.7","depth":2,"path":"binaryclass/news20_rf.md","ref":"binaryclass/news20_rf.md","articles":[]},"dir":"ltr"},"config":{"plugins":["theme-api","edit-link","github","splitter","etoc","callouts","toggle-chapters","anchorjs","codeblock-filename","expandable-chapters","multipart","codeblock-filename","katex","emphasize","localized-footer"],"styles":{"website":"styles/website.css","pdf":"styles/pdf.css","epub":"styles/epub.css","mobi":"styles/mobi.css","ebook":"styles/ebook.css","print":"styles/print.css"},"pluginsConfig":{"emphasize":{},"callouts":{},"etoc":{"h2lb":3,"header":1,"maxdepth":3,"mindepth":1,"notoc":true},"github":{"url":"https://github.com/apache/incubator-hivemall/"},"splitter":{},"search":{},"downloadpdf":{"base":"https://github.com/apache/incubator-hivemall/docs/gitbook","label":"PDF","multilingual":false},"multipart":{},"localized-footer":{"filename":"FOOTER.md","hline":"true"},"lunr":{"maxIndexSize":1000000,"ignoreSpecialCharacters":false},"katex":{},"fontsettings":{"theme":"white","family":"sans","size":2,"font":"sans"},"highlight":{},"codeblock-filename":{},"theme-api":{"languages":[],"split":false,"theme":"dark"},"sharing":{"facebook":true,"twitter":true,"google":false,"weibo":false,"instapaper":false,"vk":false,"all":["facebook","google","twitter","weibo","instapaper"]},"edit-link":{"label":"Edit","base":"https://github.com/apache/incubator-hivemall/tree/master/docs/gitbook"},"theme-default":{"styles":{"website":"styles/website.css","pdf":"styles/pdf.css","epub":"styles/epub.css","mobi":"styles/mobi.css","ebook":"styles/ebook.css","print":"styles/print.css"},"showLevel":true},"anchorjs":{"selector":"h1,h2,h3,*:not(.callout) > h4,h5"},"toggle-chapters":{},"expandable-chapters":{}},"theme":"default","pdf":{"pageNumbers":true,"fontSize":12,"fontFamily":"Arial","paperSize":"a4","chapterMark":"pagebreak","pageBreaksBefore":"/","margin":{"right":62,"left":62,"top":56,"bottom":56}},"structure":{"langs":"LANGS.md","readme":"README.md","glossary":"GLOSSARY.md","summary":"SUMMARY.md"},"variables":{},"title":"Hivemall User Manual","links":{"sidebar":{"<i class=\"fa fa-home\"></i> Home":"https://hivemall.incubator.apache.org/"}},"gitbook":"3.x.x","description":"User Manual for Apache Hivemall"},"file":{"path":"binaryclass/news20b_xgboost.md","mtime":"2021-04-22T11:42:38.110Z","type":"markdown"},"gitbook":{"version":"3.2.3","time":"2021-04-22T11:56:59.644Z"},"basePath":"..","book":{"language":""}}); |
| }); |
| </script> |
| </div> |
| |
| |
| <script src="../gitbook/gitbook.js"></script> |
| <script src="../gitbook/theme.js"></script> |
| |
| |
| <script src="../gitbook/gitbook-plugin-edit-link/plugin.js"></script> |
| |
| |
| |
| <script src="../gitbook/gitbook-plugin-github/plugin.js"></script> |
| |
| |
| |
| <script src="../gitbook/gitbook-plugin-splitter/splitter.js"></script> |
| |
| |
| |
| <script src="../gitbook/gitbook-plugin-etoc/plugin.js"></script> |
| |
| |
| |
| <script src="../gitbook/gitbook-plugin-toggle-chapters/toggle.js"></script> |
| |
| |
| |
| <script src="../gitbook/gitbook-plugin-anchorjs/anchor.min.js"></script> |
| |
| |
| |
| <script src="../gitbook/gitbook-plugin-anchorjs/anchor-style.js"></script> |
| |
| |
| |
| <script src="../gitbook/gitbook-plugin-expandable-chapters/expandable-chapters.js"></script> |
| |
| |
| |
| <script src="../gitbook/gitbook-plugin-search/search-engine.js"></script> |
| |
| |
| |
| <script src="../gitbook/gitbook-plugin-search/search.js"></script> |
| |
| |
| |
| <script src="../gitbook/gitbook-plugin-lunr/lunr.min.js"></script> |
| |
| |
| |
| <script src="../gitbook/gitbook-plugin-lunr/search-lunr.js"></script> |
| |
| |
| |
| <script src="../gitbook/gitbook-plugin-sharing/buttons.js"></script> |
| |
| |
| |
| <script src="../gitbook/gitbook-plugin-fontsettings/fontsettings.js"></script> |
| |
| |
| |
| <script src="../gitbook/gitbook-plugin-theme-api/theme-api.js"></script> |
| |
| |
| |
| </body> |
| </html> |
| |