http://git-wip-us.apache.org/repos/asf/incubator-hivemall-site/blob/30afb6e4/userguide/getting_started/input-format.html
----------------------------------------------------------------------
diff --git a/userguide/getting_started/input-format.html
b/userguide/getting_started/input-format.html
new file mode 100644
index 0000000..be88c0e
--- /dev/null
+++ b/userguide/getting_started/input-format.html
@@ -0,0 +1,1905 @@
+
+<!DOCTYPE HTML>
+<html lang="" >
+ <head>
+ <meta charset="UTF-8">
+ <meta content="text/html; charset=utf-8" http-equiv="Content-Type">
+ <title>Input Format · Hivemall User Manual</title>
+ <meta http-equiv="X-UA-Compatible" content="IE=edge" />
+ <meta name="description" content="">
+ <meta name="generator" content="GitBook 3.2.2">
+
+
+
+
+ <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="../tips/" />
+
+
+ <link rel="prev" href="permanent-functions.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="http://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="./">
+
+ <a href="./">
+
+
+ <b>1.2.</b>
+
+ Getting Started
+
+ </a>
+
+
+
+ <ul class="articles">
+
+
+ <li class="chapter " data-level="1.2.1" data-path="installation.html">
+
+ <a href="installation.html">
+
+
+ <b>1.2.1.</b>
+
+ Installation
+
+ </a>
+
+
+
+ </li>
+
+ <li class="chapter " data-level="1.2.2"
data-path="permanent-functions.html">
+
+ <a href="permanent-functions.html">
+
+
+ <b>1.2.2.</b>
+
+ Install as permanent functions
+
+ </a>
+
+
+
+ </li>
+
+ <li class="chapter active" data-level="1.2.3"
data-path="input-format.html">
+
+ <a href="input-format.html">
+
+
+ <b>1.2.3.</b>
+
+ Input Format
+
+ </a>
+
+
+
+ </li>
+
+
+ </ul>
+
+ </li>
+
+ <li class="chapter " data-level="1.3" data-path="../tips/">
+
+ <a href="../tips/">
+
+
+ <b>1.3.</b>
+
+ Tips for Effective Hivemall
+
+ </a>
+
+
+
+ <ul class="articles">
+
+
+ <li class="chapter " data-level="1.3.1"
data-path="../tips/addbias.html">
+
+ <a href="../tips/addbias.html">
+
+
+ <b>1.3.1.</b>
+
+ Explicit addBias() for better prediction
+
+ </a>
+
+
+
+ </li>
+
+ <li class="chapter " data-level="1.3.2"
data-path="../tips/rand_amplify.html">
+
+ <a href="../tips/rand_amplify.html">
+
+
+ <b>1.3.2.</b>
+
+ Use rand_amplify() to better prediction results
+
+ </a>
+
+
+
+ </li>
+
+ <li class="chapter " data-level="1.3.3"
data-path="../tips/rt_prediction.html">
+
+ <a href="../tips/rt_prediction.html">
+
+
+ <b>1.3.3.</b>
+
+ Real-time Prediction on RDBMS
+
+ </a>
+
+
+
+ </li>
+
+ <li class="chapter " data-level="1.3.4"
data-path="../tips/ensemble_learning.html">
+
+ <a href="../tips/ensemble_learning.html">
+
+
+ <b>1.3.4.</b>
+
+ Ensemble learning for stable prediction
+
+ </a>
+
+
+
+ </li>
+
+ <li class="chapter " data-level="1.3.5"
data-path="../tips/mixserver.html">
+
+ <a href="../tips/mixserver.html">
+
+
+ <b>1.3.5.</b>
+
+ Mixing models for a better prediction convergence (MIX
server)
+
+ </a>
+
+
+
+ </li>
+
+ <li class="chapter " data-level="1.3.6" data-path="../tips/emr.html">
+
+ <a href="../tips/emr.html">
+
+
+ <b>1.3.6.</b>
+
+ Run Hivemall on Amazon Elastic MapReduce
+
+ </a>
+
+
+
+ </li>
+
+
+ </ul>
+
+ </li>
+
+ <li class="chapter " data-level="1.4"
data-path="../tips/general_tips.html">
+
+ <a href="../tips/general_tips.html">
+
+
+ <b>1.4.</b>
+
+ General Hive/Hadoop tips
+
+ </a>
+
+
+
+ <ul class="articles">
+
+
+ <li class="chapter " data-level="1.4.1" data-path="../tips/rowid.html">
+
+ <a href="../tips/rowid.html">
+
+
+ <b>1.4.1.</b>
+
+ Adding rowid for each row
+
+ </a>
+
+
+
+ </li>
+
+ <li class="chapter " data-level="1.4.2"
data-path="../tips/hadoop_tuning.html">
+
+ <a href="../tips/hadoop_tuning.html">
+
+
+ <b>1.4.2.</b>
+
+ Hadoop tuning for Hivemall
+
+ </a>
+
+
+
+ </li>
+
+
+ </ul>
+
+ </li>
+
+ <li class="chapter " data-level="1.5" data-path="../troubleshooting/">
+
+ <a href="../troubleshooting/">
+
+
+ <b>1.5.</b>
+
+ Troubleshooting
+
+ </a>
+
+
+
+ <ul class="articles">
+
+
+ <li class="chapter " data-level="1.5.1"
data-path="../troubleshooting/oom.html">
+
+ <a href="../troubleshooting/oom.html">
+
+
+ <b>1.5.1.</b>
+
+ OutOfMemoryError in training
+
+ </a>
+
+
+
+ </li>
+
+ <li class="chapter " data-level="1.5.2"
data-path="../troubleshooting/mapjoin_task_error.html">
+
+ <a href="../troubleshooting/mapjoin_task_error.html">
+
+
+ <b>1.5.2.</b>
+
+ SemanticException Generate Map Join Task Error: Cannot
serialize object
+
+ </a>
+
+
+
+ </li>
+
+ <li class="chapter " data-level="1.5.3"
data-path="../troubleshooting/asterisk.html">
+
+ <a href="../troubleshooting/asterisk.html">
+
+
+ <b>1.5.3.</b>
+
+ Asterisk argument for UDTF does not work
+
+ </a>
+
+
+
+ </li>
+
+ <li class="chapter " data-level="1.5.4"
data-path="../troubleshooting/num_mappers.html">
+
+ <a href="../troubleshooting/num_mappers.html">
+
+
+ <b>1.5.4.</b>
+
+ The number of mappers is less than input splits in Hadoop
2.x
+
+ </a>
+
+
+
+ </li>
+
+ <li class="chapter " data-level="1.5.5"
data-path="../troubleshooting/mapjoin_classcastex.html">
+
+ <a href="../troubleshooting/mapjoin_classcastex.html">
+
+
+ <b>1.5.5.</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>
+
+ English/Japanese Text Tokenizer
+
+ </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/tfidf.html">
+
+ <a href="../ft_engineering/tfidf.html">
+
+
+ <b>3.3.</b>
+
+ TF-IDF calculation
+
+ </a>
+
+
+
+ </li>
+
+ <li class="chapter " data-level="3.4"
data-path="../ft_engineering/ft_trans.html">
+
+ <a href="../ft_engineering/ft_trans.html">
+
+
+ <b>3.4.</b>
+
+ FEATURE TRANSFORMATION
+
+ </a>
+
+
+
+ <ul class="articles">
+
+
+ <li class="chapter " data-level="3.4.1"
data-path="../ft_engineering/vectorizer.html">
+
+ <a href="../ft_engineering/vectorizer.html">
+
+
+ <b>3.4.1.</b>
+
+ Vectorize Features
+
+ </a>
+
+
+
+ </li>
+
+ <li class="chapter " data-level="3.4.2"
data-path="../ft_engineering/quantify.html">
+
+ <a href="../ft_engineering/quantify.html">
+
+
+ <b>3.4.2.</b>
+
+ Quantify non-number features
+
+ </a>
+
+
+
+ </li>
+
+
+ </ul>
+
+ </li>
+
+
+
+
+ <li class="header">Part IV - Evaluation</li>
+
+
+
+ <li class="chapter " data-level="4.1"
data-path="../eval/stat_eval.html">
+
+ <a href="../eval/stat_eval.html">
+
+
+ <b>4.1.</b>
+
+ Statistical evaluation of a prediction model
+
+ </a>
+
+
+
+ </li>
+
+ <li class="chapter " data-level="4.2" data-path="../eval/datagen.html">
+
+ <a href="../eval/datagen.html">
+
+
+ <b>4.2.</b>
+
+ Data Generation
+
+ </a>
+
+
+
+ <ul class="articles">
+
+
+ <li class="chapter " data-level="4.2.1"
data-path="../eval/lr_datagen.html">
+
+ <a href="../eval/lr_datagen.html">
+
+
+ <b>4.2.1.</b>
+
+ Logistic Regression data generation
+
+ </a>
+
+
+
+ </li>
+
+
+ </ul>
+
+ </li>
+
+
+
+
+ <li class="header">Part V - Binary classification</li>
+
+
+
+ <li class="chapter " data-level="5.1"
data-path="../binaryclass/a9a.html">
+
+ <a href="../binaryclass/a9a.html">
+
+
+ <b>5.1.</b>
+
+ a9a Tutorial
+
+ </a>
+
+
+
+ <ul class="articles">
+
+
+ <li class="chapter " data-level="5.1.1"
data-path="../binaryclass/a9a_dataset.html">
+
+ <a href="../binaryclass/a9a_dataset.html">
+
+
+ <b>5.1.1.</b>
+
+ Data preparation
+
+ </a>
+
+
+
+ </li>
+
+ <li class="chapter " data-level="5.1.2"
data-path="../binaryclass/a9a_lr.html">
+
+ <a href="../binaryclass/a9a_lr.html">
+
+
+ <b>5.1.2.</b>
+
+ Logistic Regression
+
+ </a>
+
+
+
+ </li>
+
+ <li class="chapter " data-level="5.1.3"
data-path="../binaryclass/a9a_minibatch.html">
+
+ <a href="../binaryclass/a9a_minibatch.html">
+
+
+ <b>5.1.3.</b>
+
+ Mini-batch Gradient Descent
+
+ </a>
+
+
+
+ </li>
+
+
+ </ul>
+
+ </li>
+
+ <li class="chapter " data-level="5.2"
data-path="../binaryclass/news20.html">
+
+ <a href="../binaryclass/news20.html">
+
+
+ <b>5.2.</b>
+
+ News20 Tutorial
+
+ </a>
+
+
+
+ <ul class="articles">
+
+
+ <li class="chapter " data-level="5.2.1"
data-path="../binaryclass/news20_dataset.html">
+
+ <a href="../binaryclass/news20_dataset.html">
+
+
+ <b>5.2.1.</b>
+
+ Data preparation
+
+ </a>
+
+
+
+ </li>
+
+ <li class="chapter " data-level="5.2.2"
data-path="../binaryclass/news20_pa.html">
+
+ <a href="../binaryclass/news20_pa.html">
+
+
+ <b>5.2.2.</b>
+
+ Perceptron, Passive Aggressive
+
+ </a>
+
+
+
+ </li>
+
+ <li class="chapter " data-level="5.2.3"
data-path="../binaryclass/news20_scw.html">
+
+ <a href="../binaryclass/news20_scw.html">
+
+
+ <b>5.2.3.</b>
+
+ CW, AROW, SCW
+
+ </a>
+
+
+
+ </li>
+
+ <li class="chapter " data-level="5.2.4"
data-path="../binaryclass/news20_adagrad.html">
+
+ <a href="../binaryclass/news20_adagrad.html">
+
+
+ <b>5.2.4.</b>
+
+ AdaGradRDA, AdaGrad, AdaDelta
+
+ </a>
+
+
+
+ </li>
+
+
+ </ul>
+
+ </li>
+
+ <li class="chapter " data-level="5.3"
data-path="../binaryclass/kdd2010a.html">
+
+ <a href="../binaryclass/kdd2010a.html">
+
+
+ <b>5.3.</b>
+
+ KDD2010a Tutorial
+
+ </a>
+
+
+
+ <ul class="articles">
+
+
+ <li class="chapter " data-level="5.3.1"
data-path="../binaryclass/kdd2010a_dataset.html">
+
+ <a href="../binaryclass/kdd2010a_dataset.html">
+
+
+ <b>5.3.1.</b>
+
+ Data preparation
+
+ </a>
+
+
+
+ </li>
+
+ <li class="chapter " data-level="5.3.2"
data-path="../binaryclass/kdd2010a_scw.html">
+
+ <a href="../binaryclass/kdd2010a_scw.html">
+
+
+ <b>5.3.2.</b>
+
+ PA, CW, AROW, SCW
+
+ </a>
+
+
+
+ </li>
+
+
+ </ul>
+
+ </li>
+
+ <li class="chapter " data-level="5.4"
data-path="../binaryclass/kdd2010b.html">
+
+ <a href="../binaryclass/kdd2010b.html">
+
+
+ <b>5.4.</b>
+
+ KDD2010b Tutorial
+
+ </a>
+
+
+
+ <ul class="articles">
+
+
+ <li class="chapter " data-level="5.4.1"
data-path="../binaryclass/kdd2010b_dataset.html">
+
+ <a href="../binaryclass/kdd2010b_dataset.html">
+
+
+ <b>5.4.1.</b>
+
+ Data preparation
+
+ </a>
+
+
+
+ </li>
+
+ <li class="chapter " data-level="5.4.2"
data-path="../binaryclass/kdd2010b_arow.html">
+
+ <a href="../binaryclass/kdd2010b_arow.html">
+
+
+ <b>5.4.2.</b>
+
+ AROW
+
+ </a>
+
+
+
+ </li>
+
+
+ </ul>
+
+ </li>
+
+ <li class="chapter " data-level="5.5"
data-path="../binaryclass/webspam.html">
+
+ <a href="../binaryclass/webspam.html">
+
+
+ <b>5.5.</b>
+
+ Webspam Tutorial
+
+ </a>
+
+
+
+ <ul class="articles">
+
+
+ <li class="chapter " data-level="5.5.1"
data-path="../binaryclass/webspam_dataset.html">
+
+ <a href="../binaryclass/webspam_dataset.html">
+
+
+ <b>5.5.1.</b>
+
+ Data pareparation
+
+ </a>
+
+
+
+ </li>
+
+ <li class="chapter " data-level="5.5.2"
data-path="../binaryclass/webspam_scw.html">
+
+ <a href="../binaryclass/webspam_scw.html">
+
+
+ <b>5.5.2.</b>
+
+ PA1, AROW, SCW
+
+ </a>
+
+
+
+ </li>
+
+
+ </ul>
+
+ </li>
+
+
+
+
+ <li class="header">Part VI - Multiclass classification</li>
+
+
+
+ <li class="chapter " data-level="6.1"
data-path="../multiclass/news20.html">
+
+ <a href="../multiclass/news20.html">
+
+
+ <b>6.1.</b>
+
+ News20 Multiclass Tutorial
+
+ </a>
+
+
+
+ <ul class="articles">
+
+
+ <li class="chapter " data-level="6.1.1"
data-path="../multiclass/news20_dataset.html">
+
+ <a href="../multiclass/news20_dataset.html">
+
+
+ <b>6.1.1.</b>
+
+ Data preparation
+
+ </a>
+
+
+
+ </li>
+
+ <li class="chapter " data-level="6.1.2"
data-path="../multiclass/news20_one-vs-the-rest_dataset.html">
+
+ <a href="../multiclass/news20_one-vs-the-rest_dataset.html">
+
+
+ <b>6.1.2.</b>
+
+ Data preparation for one-vs-the-rest classifiers
+
+ </a>
+
+
+
+ </li>
+
+ <li class="chapter " data-level="6.1.3"
data-path="../multiclass/news20_pa.html">
+
+ <a href="../multiclass/news20_pa.html">
+
+
+ <b>6.1.3.</b>
+
+ PA
+
+ </a>
+
+
+
+ </li>
+
+ <li class="chapter " data-level="6.1.4"
data-path="../multiclass/news20_scw.html">
+
+ <a href="../multiclass/news20_scw.html">
+
+
+ <b>6.1.4.</b>
+
+ CW, AROW, SCW
+
+ </a>
+
+
+
+ </li>
+
+ <li class="chapter " data-level="6.1.5"
data-path="../multiclass/news20_ensemble.html">
+
+ <a href="../multiclass/news20_ensemble.html">
+
+
+ <b>6.1.5.</b>
+
+ Ensemble learning
+
+ </a>
+
+
+
+ </li>
+
+ <li class="chapter " data-level="6.1.6"
data-path="../multiclass/news20_one-vs-the-rest.html">
+
+ <a href="../multiclass/news20_one-vs-the-rest.html">
+
+
+ <b>6.1.6.</b>
+
+ one-vs-the-rest classifier
+
+ </a>
+
+
+
+ </li>
+
+
+ </ul>
+
+ </li>
+
+ <li class="chapter " data-level="6.2"
data-path="../multiclass/iris.html">
+
+ <a href="../multiclass/iris.html">
+
+
+ <b>6.2.</b>
+
+ Iris Tutorial
+
+ </a>
+
+
+
+ <ul class="articles">
+
+
+ <li class="chapter " data-level="6.2.1"
data-path="../multiclass/iris_dataset.html">
+
+ <a href="../multiclass/iris_dataset.html">
+
+
+ <b>6.2.1.</b>
+
+ Data preparation
+
+ </a>
+
+
+
+ </li>
+
+ <li class="chapter " data-level="6.2.2"
data-path="../multiclass/iris_scw.html">
+
+ <a href="../multiclass/iris_scw.html">
+
+
+ <b>6.2.2.</b>
+
+ SCW
+
+ </a>
+
+
+
+ </li>
+
+ <li class="chapter " data-level="6.2.3"
data-path="../multiclass/iris_randomforest.html">
+
+ <a href="../multiclass/iris_randomforest.html">
+
+
+ <b>6.2.3.</b>
+
+ RandomForest
+
+ </a>
+
+
+
+ </li>
+
+
+ </ul>
+
+ </li>
+
+
+
+
+ <li class="header">Part VII - Regression</li>
+
+
+
+ <li class="chapter " data-level="7.1"
data-path="../regression/e2006.html">
+
+ <a href="../regression/e2006.html">
+
+
+ <b>7.1.</b>
+
+ E2006-tfidf regression Tutorial
+
+ </a>
+
+
+
+ <ul class="articles">
+
+
+ <li class="chapter " data-level="7.1.1"
data-path="../regression/e2006_dataset.html">
+
+ <a href="../regression/e2006_dataset.html">
+
+
+ <b>7.1.1.</b>
+
+ Data preparation
+
+ </a>
+
+
+
+ </li>
+
+ <li class="chapter " data-level="7.1.2"
data-path="../regression/e2006_arow.html">
+
+ <a href="../regression/e2006_arow.html">
+
+
+ <b>7.1.2.</b>
+
+ Passive Aggressive, AROW
+
+ </a>
+
+
+
+ </li>
+
+
+ </ul>
+
+ </li>
+
+ <li class="chapter " data-level="7.2"
data-path="../regression/kddcup12tr2.html">
+
+ <a href="../regression/kddcup12tr2.html">
+
+
+ <b>7.2.</b>
+
+ KDDCup 2012 track 2 CTR prediction Tutorial
+
+ </a>
+
+
+
+ <ul class="articles">
+
+
+ <li class="chapter " data-level="7.2.1"
data-path="../regression/kddcup12tr2_dataset.html">
+
+ <a href="../regression/kddcup12tr2_dataset.html">
+
+
+ <b>7.2.1.</b>
+
+ Data preparation
+
+ </a>
+
+
+
+ </li>
+
+ <li class="chapter " data-level="7.2.2"
data-path="../regression/kddcup12tr2_lr.html">
+
+ <a href="../regression/kddcup12tr2_lr.html">
+
+
+ <b>7.2.2.</b>
+
+ Logistic Regression, Passive Aggressive
+
+ </a>
+
+
+
+ </li>
+
+ <li class="chapter " data-level="7.2.3"
data-path="../regression/kddcup12tr2_lr_amplify.html">
+
+ <a href="../regression/kddcup12tr2_lr_amplify.html">
+
+
+ <b>7.2.3.</b>
+
+ Logistic Regression with Amplifier
+
+ </a>
+
+
+
+ </li>
+
+ <li class="chapter " data-level="7.2.4"
data-path="../regression/kddcup12tr2_adagrad.html">
+
+ <a href="../regression/kddcup12tr2_adagrad.html">
+
+
+ <b>7.2.4.</b>
+
+ AdaGrad, AdaDelta
+
+ </a>
+
+
+
+ </li>
+
+
+ </ul>
+
+ </li>
+
+
+
+
+ <li class="header">Part VIII - Recommendation</li>
+
+
+
+ <li class="chapter " data-level="8.1" data-path="../recommend/cf.html">
+
+ <a href="../recommend/cf.html">
+
+
+ <b>8.1.</b>
+
+ Collaborative Filtering
+
+ </a>
+
+
+
+ <ul class="articles">
+
+
+ <li class="chapter " data-level="8.1.1"
data-path="../recommend/item_based_cf.html">
+
+ <a href="../recommend/item_based_cf.html">
+
+
+ <b>8.1.1.</b>
+
+ Item-based Collaborative Filtering
+
+ </a>
+
+
+
+ </li>
+
+
+ </ul>
+
+ </li>
+
+ <li class="chapter " data-level="8.2"
data-path="../recommend/news20.html">
+
+ <a href="../recommend/news20.html">
+
+
+ <b>8.2.</b>
+
+ News20 related article recommendation Tutorial
+
+ </a>
+
+
+
+ <ul class="articles">
+
+
+ <li class="chapter " data-level="8.2.1"
data-path="../multiclass/news20_dataset.html">
+
+ <a href="../multiclass/news20_dataset.html">
+
+
+ <b>8.2.1.</b>
+
+ Data preparation
+
+ </a>
+
+
+
+ </li>
+
+ <li class="chapter " data-level="8.2.2"
data-path="../recommend/news20_jaccard.html">
+
+ <a href="../recommend/news20_jaccard.html">
+
+
+ <b>8.2.2.</b>
+
+ LSH/Minhash and Jaccard Similarity
+
+ </a>
+
+
+
+ </li>
+
+ <li class="chapter " data-level="8.2.3"
data-path="../recommend/news20_knn.html">
+
+ <a href="../recommend/news20_knn.html">
+
+
+ <b>8.2.3.</b>
+
+ LSH/Minhash and Brute-Force Search
+
+ </a>
+
+
+
+ </li>
+
+ <li class="chapter " data-level="8.2.4"
data-path="../recommend/news20_bbit_minhash.html">
+
+ <a href="../recommend/news20_bbit_minhash.html">
+
+
+ <b>8.2.4.</b>
+
+ kNN search using b-Bits Minhash
+
+ </a>
+
+
+
+ </li>
+
+
+ </ul>
+
+ </li>
+
+ <li class="chapter " data-level="8.3"
data-path="../recommend/movielens.html">
+
+ <a href="../recommend/movielens.html">
+
+
+ <b>8.3.</b>
+
+ MovieLens movie recommendation Tutorial
+
+ </a>
+
+
+
+ <ul class="articles">
+
+
+ <li class="chapter " data-level="8.3.1"
data-path="../recommend/movielens_dataset.html">
+
+ <a href="../recommend/movielens_dataset.html">
+
+
+ <b>8.3.1.</b>
+
+ Data preparation
+
+ </a>
+
+
+
+ </li>
+
+ <li class="chapter " data-level="8.3.2"
data-path="../recommend/movielens_mf.html">
+
+ <a href="../recommend/movielens_mf.html">
+
+
+ <b>8.3.2.</b>
+
+ Matrix Factorization
+
+ </a>
+
+
+
+ </li>
+
+ <li class="chapter " data-level="8.3.3"
data-path="../recommend/movielens_fm.html">
+
+ <a href="../recommend/movielens_fm.html">
+
+
+ <b>8.3.3.</b>
+
+ Factorization Machine
+
+ </a>
+
+
+
+ </li>
+
+ <li class="chapter " data-level="8.3.4"
data-path="../recommend/movielens_cv.html">
+
+ <a href="../recommend/movielens_cv.html">
+
+
+ <b>8.3.4.</b>
+
+ 10-fold Cross Validation (Matrix Factorization)
+
+ </a>
+
+
+
+ </li>
+
+
+ </ul>
+
+ </li>
+
+
+
+
+ <li class="header">Part IX - Anomaly Detection</li>
+
+
+
+ <li class="chapter " data-level="9.1" data-path="../anomaly/lof.html">
+
+ <a href="../anomaly/lof.html">
+
+
+ <b>9.1.</b>
+
+ Outlier Detection using Local Outlier Factor (LOF)
+
+ </a>
+
+
+
+ </li>
+
+
+
+
+ <li class="header">Part X - External References</li>
+
+
+
+ <li class="chapter " data-level="10.1" >
+
+ <a target="_blank"
href="https://github.com/maropu/hivemall-spark";>
+
+
+ <b>10.1.</b>
+
+ Hivemall on Apache Spark
+
+ </a>
+
+
+
+ </li>
+
+ <li class="chapter " data-level="10.2" >
+
+ <a target="_blank"
href="https://github.com/daijyc/hivemall/wiki/PigHome";>
+
+
+ <b>10.2.</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=".." >Input Format</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">
+
+ <p>This page explains the input format of
training data in Hivemall.
+Here, we use <a
href="http://en.wikipedia.org/wiki/Extended_Backus%E2%80%93Naur_Form";
target="_blank">EBNF</a>-like notation for describing the format.</p>
+<!-- toc --><div id="toc" class="toc">
+
+<ul>
+<li><a href="#input-format-for-classification">Input Format for
Classification</a></li>
+<li><a href="#features-format-for-classification-and-regression">Features
format (for classification and regression)</a><ul>
+<li><a href="#quantitative-and-categorical-variables">Quantitative and
Categorical variables</a></li>
+<li><a href="#biasdummy-variable-in-features">Bias/Dummy Variable in
features</a></li>
+<li><a href="#feature-hashing">Feature hashing</a></li>
+<li><a href="#feature-normalization">Feature Normalization</a></li>
+</ul>
+</li>
+<li><a href="#label-format-in-binary-classification">Label format in Binary
Classification</a></li>
+<li><a href="#label-format-in-multi-class-classification">Label format in
Multi-class Classification</a></li>
+<li><a href="#input-format-in-regression">Input format in Regression</a><ul>
+<li><a href="#target-in-logistic-regression">Target in Logistic
Regression</a></li>
+</ul>
+</li>
+<li><a href="#helper-functions">Helper functions</a><ul>
+<li><a href="#quantitative-features">Quantitative Features</a></li>
+<li><a href="#categorical-features">Categorical Features</a></li>
+<li><a href="#preparing-training-data-table">Preparing training data
table</a></li>
+</ul>
+</li>
+</ul>
+
+</div><!-- tocstop -->
+<h1 id="input-format-for-classification">Input Format for Classification</h1>
+<p>The classifiers of Hivemall takes 2 (or 3) arguments: <em>features</em>,
<em>label</em>, and <em>options</em> (a.k.a. <a
href="http://en.wikipedia.org/wiki/Hyperparameter";
target="_blank">hyperparameters</a>). The first two arguments of training
functions (e.g., <a
href="https://github.com/myui/hivemall/wiki/a9a-binary-classification-(logistic-regression"
target="_blank">logress</a>) and <a
href="https://github.com/myui/hivemall/wiki/news20-binary-classification-%232-(CW,-AROW,-SCW"
target="_blank">train_scw</a>)) represents training examples. </p>
+<p>In Statistics, <em>features</em> and <em>label</em> are called <a
href="http://www.oswego.edu/~srp/stats/variable_types.htm";
target="_blank">Explanatory variable and Response Variable</a>,
respectively.</p>
+<h1 id="features-format-for-classification-and-regression">Features format
(for classification and regression)</h1>
+<p>The format of <em>features</em> is common between (binary and multi-class)
classification and regression.
+Hivemall accepts ARRAY<INT|BIGINT|TEXT> for the type of
<em>features</em> column.</p>
+<p>Hivemall uses a <em>sparse</em> data format (cf. <a
href="http://netlib.org/linalg/html_templates/node91.html";
target="_blank">Compressed Row Storage</a>) which is similar to <a
href="http://stackoverflow.com/questions/12112558/read-write-data-in-libsvm-format";
target="_blank">LIBSVM</a> and <a
href="https://github.com/JohnLangford/vowpal_wabbit/wiki/Input-format";
target="_blank">Vowpal Wabbit</a>.</p>
+<p>The format of each feature in an array is as follows:</p>
+<pre><code>feature ::= <index>:<weight> or <index>
+</code></pre><p>Each element of <em>index</em> or <em>weight</em> then accepts
the following format:</p>
+<pre><code>index ::= <INT | BIGINT | TEXT>
+weight ::= <FLOAT>
+</code></pre><p>The <em>index</em> are usually a number (INT or BIGINT)
starting from 1.
+Here is an instance of a features.</p>
+<pre><code>10:3.4 123:0.5 34567:0.231
+</code></pre><p><em>Note:</em> As mentioned later, <em>index</em>
"0" is reserved for a <a
href="https://github.com/myui/hivemall/wiki/Using-explicit-addBias("
target="_blank">Bias/Dummy variable</a>-for-a-better-prediction).</p>
+<p>In addition to numbers, you can use a TEXT value for an index. For example,
you can use array("height:1.5", "length:2.0") for the
features.</p>
+<pre><code>"height:1.5" "length:2.0"
+</code></pre><h2 id="quantitative-and-categorical-variables">Quantitative and
Categorical variables</h2>
+<p>A <a href="http://www.oswego.edu/~srp/stats/variable_types.htm";
target="_blank">quantitative variable</a> must have an <em>index</em> entry.</p>
+<p>Hivemall (v0.3.1 or later) provides <em>add_feature_index</em> function
which is useful for adding indexes to quantitative variables. </p>
+<pre><code class="lang-sql"><span class="hljs-keyword">select</span>
add_feature_index(<span class="hljs-built_in">array</span>(<span
class="hljs-number">3</span>,<span class="hljs-number">4.0</span>,<span
class="hljs-number">5</span>)) <span class="hljs-keyword">from</span> dual;
+</code></pre>
+<blockquote>
+<p>["1:3.0","2:4.0","3:5.0"]</p>
+</blockquote>
+<p>You can omit specifying <em>weight</em> for each feature e.g. for <a
href="http://www.oswego.edu/~srp/stats/variable_types.htm";
target="_blank">Categorical variables</a> as follows:</p>
+<pre><code>feature ::= <index>
+</code></pre><p>Note 1.0 is used for the weight when omitting <em>weight</em>.
</p>
+<h2 id="biasdummy-variable-in-features">Bias/Dummy Variable in features</h2>
+<p>Note that "0" is reserved for a Bias variable (called dummy
variable in Statistics). </p>
+<p>The <a href="https://github.com/myui/hivemall/wiki/Using-explicit-addBias("
target="_blank">addBias</a>-for-a-better-prediction) function is Hivemall
appends "0:1.0" as an element of array in <em>features</em>.</p>
+<h2 id="feature-hashing">Feature hashing</h2>
+<p>Hivemall supports <a href="http://en.wikipedia.org/wiki/Feature_hashing";
target="_blank">feature hashing/hashing trick</a> through <a
href="https://github.com/myui/hivemall/wiki/KDDCup-2012-track-2-CTR-prediction-dataset#converting-feature-representation-by-feature-hashing";
target="_blank">mhash function</a>.</p>
+<p>The mhash function takes a feature (i.e., <em>index</em>) of TEXT format
and generates a hash number of a range from 1 to 2^24 (=16777216) by the
default setting.</p>
+<p>Feature hashing is useful where the dimension of feature vector (i.e., the
number of elements in <em>features</em>) is so large. Consider applying <a
href="(https:/github.com/myui/hivemall/wiki/KDDCup-2012-track-2-CTR-prediction-dataset#converting-feature-representation-by-feature-hashing">mhash
function</a>) when a prediction model does not fit in memory and OutOfMemory
exception happens.</p>
+<p>In general, you don't need to use mhash when the dimension of feature
vector is less than 16777216.
+If feature <em>index</em> is very long TEXT (e.g.,
"xxxxxxx-yyyyyy-weight:55.3") and uses huge memory spaces, consider
using mhash as follows:</p>
+<pre><code class="lang-sql"><span class="hljs-comment">-- feature is v0.3.2 or
before</span>
+concat(mhash(extract_feature("xxxxxxx-yyyyyy-weight:55.3")),
":", extract_weight("xxxxxxx-yyyyyy-weight:55.3"))
+
+<span class="hljs-comment">-- feature is v0.3.2-1 or later</span>
+feature(mhash(extract_feature("xxxxxxx-yyyyyy-weight:55.3")),
extract_weight("xxxxxxx-yyyyyy-weight:55.3"))
+</code></pre>
+<blockquote>
+<p>43352:55.3</p>
+</blockquote>
+<h2 id="feature-normalization">Feature Normalization</h2>
+<p>Feature (weight) normalization is important in machine learning. Please
refer <a href="https://github.com/myui/hivemall/wiki/Feature-scaling";
target="_blank">https://github.com/myui/hivemall/wiki/Feature-scaling</a> for
detail.</p>
+<hr>
+<h1 id="label-format-in-binary-classification">Label format in Binary
Classification</h1>
+<p>The <em>label</em> must be an <em>INT</em> 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><h1 id="label-format-in-multi-class-classification">Label format
in Multi-class Classification</h1>
+<p>You can used any PRIMITIVE type in the multi-class <em>label</em>. </p>
+<pre><code><label> ::= <primitive type>
+</code></pre><p>Typically, the type of label column will be INT, BIGINT, or
TEXT.</p>
+<hr>
+<h1 id="input-format-in-regression">Input format in Regression</h1>
+<p>In regression, response/predictor variable (we denote it as
<em>target</em>) is a real number.</p>
+<p>Before Hivemall v0.3, we accepts only FLOAT type for <em>target</em>.</p>
+<pre><code><target> ::= <FLOAT>
+</code></pre><p>You need to explicitly cast a double value of <em>target</em>
to float as follows:</p>
+<pre><code class="lang-sql">CAST(target as FLOAT)
+</code></pre>
+<p>On the other hand, Hivemall v0.3 or later accepts double compatible numbers
in <em>target</em>.</p>
+<pre><code><target> ::= <FLOAT | DOUBLE | INT | TINYINT | SMALLINT|
BIGINT >
+</code></pre><h2 id="target-in-logistic-regression">Target in Logistic
Regression</h2>
+<p>Logistic regression is actually a binary classification scheme while it can
produce probabilities of positive of a training example. </p>
+<p>A <em>target</em> value of a training input must be in range 0.0 to 1.0,
specifically 0.0 or 1.0.</p>
+<hr>
+<h1 id="helper-functions">Helper functions</h1>
+<pre><code class="lang-sql"><span class="hljs-comment">-- hivemall v0.3.2 and
before</span>
+<span class="hljs-keyword">select</span> <span
class="hljs-keyword">concat</span>(<span
class="hljs-string">"weight"</span>,<span
class="hljs-string">":"</span>,<span class="hljs-number">55.0</span>);
+
+<span class="hljs-comment">-- hivemall v0.3.2-1 and later</span>
+<span class="hljs-keyword">select</span> feature(<span
class="hljs-string">"weight"</span>, <span
class="hljs-number">55.0</span>);
+</code></pre>
+<blockquote>
+<p>weight:55.0</p>
+</blockquote>
+<pre><code class="lang-sql"><span class="hljs-keyword">select</span>
extract_feature(<span class="hljs-string">"weight:55.0"</span>),
extract_weight(<span class="hljs-string">"weight:55.0"</span>);
+</code></pre>
+<blockquote>
+<p>weight | 55.0</p>
+</blockquote>
+<pre><code class="lang-sql"><span class="hljs-comment">-- hivemall v0.4.0 and
later</span>
+<span class="hljs-keyword">select</span> feature_index(<span
class="hljs-built_in">array</span>(<span
class="hljs-string">"10:0.2"</span>,<span
class="hljs-string">"7:0.3"</span>,<span
class="hljs-string">"9"</span>));
+</code></pre>
+<blockquote>
+<p>[10,7,9]</p>
+</blockquote>
+<pre><code class="lang-sql"><span class="hljs-keyword">select</span>
+ convert_label(<span class="hljs-number">-1</span>), convert_label(<span
class="hljs-number">1</span>), convert_label(<span
class="hljs-number">0.0</span>f), convert_label(<span
class="hljs-number">1.0</span>f)
+<span class="hljs-keyword">from</span>
+ dual;
+</code></pre>
+<blockquote>
+<p>0.0f | 1.0f | -1 | 1</p>
+</blockquote>
+<h2 id="quantitative-features">Quantitative Features</h2>
+<p><code>array<string> quantitative_features(array<string>
featureNames, ...)</code> is a helper function to create sparse quantitative
features from a table.</p>
+<pre><code class="lang-sql"><span class="hljs-keyword">select</span>
quantitative_features(<span class="hljs-built_in">array</span>(<span
class="hljs-string">"apple"</span>,<span
class="hljs-string">"value"</span>),<span
class="hljs-number">1</span>,<span class="hljs-number">120.3</span>);
+</code></pre>
+<blockquote>
+<p>["apple:1.0","value:120.3"]</p>
+</blockquote>
+<h2 id="categorical-features">Categorical Features</h2>
+<p><code>array<string> categorical_features(array<string>
featureNames, ...)</code> is a helper function to create sparse categorical
features from a table.</p>
+<pre><code class="lang-sql"><span class="hljs-keyword">select</span>
categorical_features(
+ <span class="hljs-built_in">array</span>(<span
class="hljs-string">"is_cat"</span>,<span
class="hljs-string">"is_dog"</span>,<span
class="hljs-string">"is_lion"</span>,<span
class="hljs-string">"is_pengin"</span>,<span
class="hljs-string">"species"</span>),
+ <span class="hljs-number">1</span>, <span class="hljs-number">0</span>,
<span class="hljs-number">1.0</span>, <span class="hljs-literal">true</span>,
<span class="hljs-string">"dog"</span>
+);
+</code></pre>
+<blockquote>
+<p>["is_cat#1","is_dog#0","is_lion#1.0","is_pengin#true","species#dog"]</p>
+</blockquote>
+<h2 id="preparing-training-data-table">Preparing training data table</h2>
+<p>You can create a training data table as follows:</p>
+<pre><code class="lang-sql"><span class="hljs-keyword">select</span>
+ <span class="hljs-keyword">rowid</span>() <span
class="hljs-keyword">as</span> <span class="hljs-keyword">rowid</span>,
+ concat_array(
+ <span class="hljs-built_in">array</span>(<span
class="hljs-string">"bias:1.0"</span>),
+ categorical_features(
+ <span class="hljs-built_in">array</span>(<span
class="hljs-string">"id"</span>, <span
class="hljs-string">"name"</span>),
+ <span class="hljs-keyword">id</span>, <span
class="hljs-keyword">name</span>
+ ),
+ quantitative_features(
+ <span class="hljs-built_in">array</span>(<span
class="hljs-string">"height"</span>, <span
class="hljs-string">"weight"</span>),
+ height, weight
+ )
+ ) <span class="hljs-keyword">as</span> features,
+ click_or_not <span class="hljs-keyword">as</span> label
+<span class="hljs-keyword">from</span>
+ <span class="hljs-keyword">table</span>;
+</code></pre>
+
+
+ </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":"Input
Format","level":"1.2.3","depth":2,"next":{"title":"Tips for Effective
Hivemall","level":"1.3","depth":1,"path":"tips/README.md","ref":"tips/README.md","articles":[{"title":"Explicit
addBias() for better
prediction","level":"1.3.1","depth":2,"path":"tips/addbias.md","ref":"tips/addbias.md","articles":[]},{"title":"Use
rand_amplify() to better prediction
results","level":"1.3.2","depth":2,"path":"tips/rand_amplify.md","ref":"tips/rand_amplify.md","articles":[]},{"title":"Real-time
Prediction on
RDBMS","level":"1.3.3","depth":2,"path":"tips/rt_prediction.md","ref":"tips/rt_prediction.md","articles":[]},{"title":"Ensemble
learning for stable
prediction","level":"1.3.4","depth":2,"path":"tips/ensemble_learning.md","ref":"tips/ensemble_learning.md","articles":[]},{"title":"Mixing
models for a better prediction convergence (MIX
server)","level":"1.3.5","depth":2,"path":"tips/mixserver.md","ref":"tips/mixserver.md","articles":[]},{
"title":"Run Hivemall on Amazon Elastic
MapReduce","level":"1.3.6","depth":2,"path":"tips/emr.md","ref":"tips/emr.md","articles":[]}]},"previous":{"title":"Install
as permanent
functions","level":"1.2.2","depth":2,"path":"getting_started/permanent-functions.md","ref":"getting_started/permanent-functions.md","articles":[]},"dir":"ltr"},"config":{"plugins":["theme-api","edit-link","github","splitter","sitemap","etoc","callouts","toggle-chapters","anchorjs","codeblock-filename","expandable-chapters","multipart","codeblock-filename","katex","emphasize"],"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":{"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","labe
l":"PDF","multilingual":false},"multipart":{},"lunr":{"maxIndexSize":1000000,"ignoreSpecialCharacters":false},"katex":{},"fontsettings":{"theme":"white","family":"sans","size":2,"font":"sans"},"highlight":{},"codeblock-filename":{},"sitemap":{"hostname":"http://hivemall.incubator.apache.org/"},"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/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":"http://hivemall.incubator.apache.org/"}},"gitbook":"3.x.x","description":"User
Manual for Apache
Hivemall"},"file":{"path":"getting_started/input-format.md","mtime":"2016-10-22T16:59:05.000Z","type":"markdown"},"gitbook":{"version":"3.2.2","time":"2016-10-22T19:10:36.585Z"},"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="https://cdnjs.cloudflare.com/ajax/libs/anchor-js/3.1.1/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>
+
