You are viewing a plain text version of this content. The canonical link for it is here.
Posted to commits@hivemall.apache.org by my...@apache.org on 2016/10/22 19:23:15 UTC
[25/51] [partial] incubator-hivemall-site git commit: Added the user
guide
http://git-wip-us.apache.org/repos/asf/incubator-hivemall-site/blob/30afb6e4/userguide/ft_engineering/scaling.html
----------------------------------------------------------------------
diff --git a/userguide/ft_engineering/scaling.html b/userguide/ft_engineering/scaling.html
new file mode 100644
index 0000000..2840c5b
--- /dev/null
+++ b/userguide/ft_engineering/scaling.html
@@ -0,0 +1,1891 @@
+
+<!DOCTYPE HTML>
+<html lang="" >
+ <head>
+ <meta charset="UTF-8">
+ <meta content="text/html; charset=utf-8" http-equiv="Content-Type">
+ <title>Feature Scaling � 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="hashing.html" />
+
+
+ <link rel="prev" href="../misc/tokenizer.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="../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="../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 active" data-level="3.1" data-path="scaling.html">
+
+ <a href="scaling.html">
+
+
+ <b>3.1.</b>
+
+ Feature Scaling
+
+ </a>
+
+
+
+ </li>
+
+ <li class="chapter " data-level="3.2" data-path="hashing.html">
+
+ <a href="hashing.html">
+
+
+ <b>3.2.</b>
+
+ Feature Hashing
+
+ </a>
+
+
+
+ </li>
+
+ <li class="chapter " data-level="3.3" data-path="tfidf.html">
+
+ <a href="tfidf.html">
+
+
+ <b>3.3.</b>
+
+ TF-IDF calculation
+
+ </a>
+
+
+
+ </li>
+
+ <li class="chapter " data-level="3.4" data-path="ft_trans.html">
+
+ <a href="ft_trans.html">
+
+
+ <b>3.4.</b>
+
+ FEATURE TRANSFORMATION
+
+ </a>
+
+
+
+ <ul class="articles">
+
+
+ <li class="chapter " data-level="3.4.1" data-path="vectorizer.html">
+
+ <a href="vectorizer.html">
+
+
+ <b>3.4.1.</b>
+
+ Vectorize Features
+
+ </a>
+
+
+
+ </li>
+
+ <li class="chapter " data-level="3.4.2" data-path="quantify.html">
+
+ <a href="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=".." >Feature Scaling</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">
+
+ <h1 id="min-max-normalization">Min-Max Normalization</h1>
+<p><a href="http://en.wikipedia.org/wiki/Feature_scaling#Rescaling" target="_blank">http://en.wikipedia.org/wiki/Feature_scaling#Rescaling</a></p>
+<pre><code class="lang-sql"><span class="hljs-keyword">select</span> <span class="hljs-keyword">min</span>(target), <span class="hljs-keyword">max</span>(target)
+<span class="hljs-keyword">from</span> (
+<span class="hljs-keyword">select</span> target <span class="hljs-keyword">from</span> e2006tfidf_train
+<span class="hljs-comment">-- union all</span>
+<span class="hljs-comment">-- select target from e2006tfidf_test </span>
+) t;
+</code></pre>
+<blockquote>
+<p>-7.899578 -0.51940954</p>
+</blockquote>
+<pre><code class="lang-sql"><span class="hljs-keyword">set</span> hivevar:min_target=<span class="hljs-number">-7.899578</span>;
+<span class="hljs-keyword">set</span> hivevar:max_target=<span class="hljs-number">-0.51940954</span>;
+
+<span class="hljs-keyword">create</span> <span class="hljs-keyword">or</span> <span class="hljs-keyword">replace</span> <span class="hljs-keyword">view</span> e2006tfidf_train_scaled
+<span class="hljs-keyword">as</span>
+<span class="hljs-keyword">select</span>
+ <span class="hljs-keyword">rowid</span>,
+ rescale(target, ${min_target}, ${max_target}) <span class="hljs-keyword">as</span> target,
+ features
+<span class="hljs-keyword">from</span>
+ e2006tfidf_train;
+</code></pre>
+<h1 id="feature-scaling-by-zscore">Feature scaling by zscore</h1>
+<p><a href="http://en.wikipedia.org/wiki/Standard_score" target="_blank">http://en.wikipedia.org/wiki/Standard_score</a></p>
+<pre><code class="lang-sql"><span class="hljs-keyword">select</span> <span class="hljs-keyword">avg</span>(target), <span class="hljs-keyword">stddev_pop</span>(target)
+<span class="hljs-keyword">from</span> (
+<span class="hljs-keyword">select</span> target <span class="hljs-keyword">from</span> e2006tfidf_train
+<span class="hljs-comment">-- union all</span>
+<span class="hljs-comment">-- select target from e2006tfidf_test </span>
+) t;
+</code></pre>
+<blockquote>
+<p>-3.566241460963296 0.6278076335455348</p>
+</blockquote>
+<pre><code class="lang-sql"><span class="hljs-keyword">set</span> hivevar:mean_target=<span class="hljs-number">-3.566241460963296</span>;
+<span class="hljs-keyword">set</span> hivevar:stddev_target=<span class="hljs-number">0.6278076335455348</span>;
+
+<span class="hljs-keyword">create</span> <span class="hljs-keyword">or</span> <span class="hljs-keyword">replace</span> <span class="hljs-keyword">view</span> e2006tfidf_train_scaled
+<span class="hljs-keyword">as</span>
+<span class="hljs-keyword">select</span>
+ <span class="hljs-keyword">rowid</span>,
+ zscore(target, ${mean_target}, ${stddev_target}) <span class="hljs-keyword">as</span> target,
+ features
+<span class="hljs-keyword">from</span>
+ e2006tfidf_train;
+</code></pre>
+<h1 id="apply-normalization-to-more-complex-feature-vector">Apply Normalization to more complex feature vector</h1>
+<p>Apply normalization to the following data.</p>
+<pre><code class="lang-sql"><span class="hljs-keyword">select</span> <span class="hljs-keyword">rowid</span>, features <span class="hljs-keyword">from</span> train <span class="hljs-keyword">limit</span> <span class="hljs-number">3</span>;
+</code></pre>
+<pre><code>1 ["weight:69.613","specific_heat:129.07","reflectance:52.111"]
+2 ["weight:70.67","specific_heat:128.161","reflectance:52.446"]
+3 ["weight:72.303","specific_heat:128.45","reflectance:52.853"]
+</code></pre><p>We can create a normalized table as follows:</p>
+<pre><code class="lang-sql"><span class="hljs-keyword">create</span> <span class="hljs-keyword">table</span> train_normalized
+<span class="hljs-keyword">as</span>
+<span class="hljs-keyword">WITH</span> fv <span class="hljs-keyword">as</span> (
+<span class="hljs-keyword">select</span>
+ <span class="hljs-keyword">rowid</span>,
+ extract_feature(feature) <span class="hljs-keyword">as</span> feature,
+ extract_weight(feature) <span class="hljs-keyword">as</span> <span class="hljs-keyword">value</span>
+<span class="hljs-keyword">from</span>
+ train
+ LATERAL <span class="hljs-keyword">VIEW</span> explode(features) exploded <span class="hljs-keyword">AS</span> feature
+),
+stats <span class="hljs-keyword">as</span> (
+<span class="hljs-keyword">select</span>
+ feature,
+ <span class="hljs-comment">-- avg(value) as mean, stddev_pop(value) as stddev</span>
+ <span class="hljs-keyword">min</span>(<span class="hljs-keyword">value</span>) <span class="hljs-keyword">as</span> <span class="hljs-keyword">min</span>, <span class="hljs-keyword">max</span>(<span class="hljs-keyword">value</span>) <span class="hljs-keyword">as</span> <span class="hljs-keyword">max</span>
+<span class="hljs-keyword">from</span>
+ fv
+<span class="hljs-keyword">group</span> <span class="hljs-keyword">by</span>
+ feature
+),
+norm <span class="hljs-keyword">as</span> (
+<span class="hljs-keyword">select</span>
+ <span class="hljs-keyword">rowid</span>,
+ t1.feature,
+ <span class="hljs-comment">-- zscore(t1.value, t2.mean, t2.stddev) as zscore</span>
+ rescale(t1.<span class="hljs-keyword">value</span>, t2.<span class="hljs-keyword">min</span>, t2.<span class="hljs-keyword">max</span>) <span class="hljs-keyword">as</span> minmax
+<span class="hljs-keyword">from</span>
+ fv t1 <span class="hljs-keyword">JOIN</span>
+ stats t2 <span class="hljs-keyword">ON</span> (t1.feature = t2.feature)
+),
+norm_fv <span class="hljs-keyword">as</span> (
+<span class="hljs-keyword">select</span>
+ <span class="hljs-keyword">rowid</span>,
+ <span class="hljs-comment">-- concat(feature, ":", zscore) as feature</span>
+ <span class="hljs-comment">-- concat(feature, ":", minmax) as feature -- Before Hivemall v0.3.2-1</span>
+ feature(feature, minmax) <span class="hljs-keyword">as</span> feature <span class="hljs-comment">-- Hivemall v0.3.2-1 or later</span>
+<span class="hljs-keyword">from</span>
+ norm
+)
+<span class="hljs-keyword">select</span>
+ <span class="hljs-keyword">rowid</span>,
+ collect_list(feature) <span class="hljs-keyword">as</span> features
+<span class="hljs-keyword">from</span>
+ norm_fv
+<span class="hljs-keyword">group</span> <span class="hljs-keyword">by</span>
+ <span class="hljs-keyword">rowid</span>
+;
+</code></pre>
+<pre><code>1 ["reflectance:0.5252967","specific_heat:0.19863537","weight:0.0"]
+2 ["reflectance:0.5950446","specific_heat:0.09166764","weight:0.052084323"]
+3 ["reflectance:0.6797837","specific_heat:0.12567581","weight:0.13255163"]
+...
+</code></pre><h1 id="tips-for-using-both-min-max-and-zscore-normalization">Tips for using both min-max and zscore normalization</h1>
+<pre><code class="lang-sql">WITH quantative as (
+ <span class="hljs-keyword">select</span> <span class="hljs-keyword">id</span>, <span class="hljs-literal">true</span> <span class="hljs-keyword">as</span> minmax, <span class="hljs-string">"age"</span> <span class="hljs-keyword">as</span> feature, age <span class="hljs-keyword">as</span> <span class="hljs-keyword">value</span> <span class="hljs-keyword">from</span> train
+ <span class="hljs-keyword">union</span> all
+ <span class="hljs-keyword">select</span> <span class="hljs-keyword">id</span>, <span class="hljs-literal">false</span> <span class="hljs-keyword">as</span> minmax, <span class="hljs-string">"balance"</span> <span class="hljs-keyword">as</span> feature, balance <span class="hljs-keyword">as</span> <span class="hljs-keyword">value</span> <span class="hljs-keyword">from</span> train
+ <span class="hljs-keyword">union</span> all
+ <span class="hljs-keyword">select</span> <span class="hljs-keyword">id</span>, <span class="hljs-literal">true</span> <span class="hljs-keyword">as</span> minmax, <span class="hljs-string">"day"</span> <span class="hljs-keyword">as</span> feature, <span class="hljs-keyword">day</span> <span class="hljs-keyword">as</span> <span class="hljs-keyword">value</span> <span class="hljs-keyword">from</span> train
+ <span class="hljs-keyword">union</span> all
+ <span class="hljs-keyword">select</span> <span class="hljs-keyword">id</span>, <span class="hljs-literal">false</span> <span class="hljs-keyword">as</span> minmax, <span class="hljs-string">"duration"</span> <span class="hljs-keyword">as</span> feature, <span class="hljs-keyword">duration</span> <span class="hljs-keyword">as</span> <span class="hljs-keyword">value</span> <span class="hljs-keyword">from</span> train
+ <span class="hljs-keyword">union</span> all
+ <span class="hljs-keyword">select</span> <span class="hljs-keyword">id</span>, <span class="hljs-literal">false</span> <span class="hljs-keyword">as</span> minmax, <span class="hljs-string">"campaign"</span> <span class="hljs-keyword">as</span> feature, campaign <span class="hljs-keyword">as</span> <span class="hljs-keyword">value</span> <span class="hljs-keyword">from</span> train
+ <span class="hljs-keyword">union</span> all
+ <span class="hljs-keyword">select</span> <span class="hljs-keyword">id</span>, <span class="hljs-literal">false</span> <span class="hljs-keyword">as</span> minmax, <span class="hljs-string">"pdays"</span> <span class="hljs-keyword">as</span> feature, <span class="hljs-keyword">if</span>(pdays = <span class="hljs-number">-1</span>, <span class="hljs-number">0</span>, pdays) <span class="hljs-keyword">as</span> <span class="hljs-keyword">value</span> <span class="hljs-keyword">from</span> train
+ <span class="hljs-keyword">union</span> all
+ <span class="hljs-keyword">select</span> <span class="hljs-keyword">id</span>, <span class="hljs-literal">false</span> <span class="hljs-keyword">as</span> minmax, <span class="hljs-string">"previous"</span> <span class="hljs-keyword">as</span> feature, previous <span class="hljs-keyword">as</span> <span class="hljs-keyword">value</span> <span class="hljs-keyword">from</span> train
+),
+quantative_stats <span class="hljs-keyword">as</span> (
+<span class="hljs-keyword">select</span>
+ feature,
+ <span class="hljs-keyword">avg</span>(<span class="hljs-keyword">value</span>) <span class="hljs-keyword">as</span> mean, <span class="hljs-keyword">stddev_pop</span>(<span class="hljs-keyword">value</span>) <span class="hljs-keyword">as</span> <span class="hljs-keyword">stddev</span>,
+ <span class="hljs-keyword">min</span>(<span class="hljs-keyword">value</span>) <span class="hljs-keyword">as</span> <span class="hljs-keyword">min</span>, <span class="hljs-keyword">max</span>(<span class="hljs-keyword">value</span>) <span class="hljs-keyword">as</span> <span class="hljs-keyword">max</span>
+<span class="hljs-keyword">from</span>
+ quantative
+<span class="hljs-keyword">group</span> <span class="hljs-keyword">by</span>
+ feature
+),
+quantative_norm <span class="hljs-keyword">as</span> (
+<span class="hljs-keyword">select</span>
+ t1.<span class="hljs-keyword">id</span>,
+ collect_list(
+ feature(
+ t1.feature,
+ <span class="hljs-keyword">if</span>(t1.minmax,rescale(t1.<span class="hljs-keyword">value</span>, t2.<span class="hljs-keyword">min</span>, t2.<span class="hljs-keyword">max</span>),zscore(t1.<span class="hljs-keyword">value</span>, t2.mean, t2.<span class="hljs-keyword">stddev</span>))
+ )
+ ) <span class="hljs-keyword">as</span> features
+<span class="hljs-keyword">from</span>
+ quantative t1
+ <span class="hljs-keyword">JOIN</span> quantative_stats t2 <span class="hljs-keyword">ON</span> (t1.feature = t2.feature)
+<span class="hljs-keyword">group</span> <span class="hljs-keyword">by</span>
+ t1.<span class="hljs-keyword">id</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":"Feature Scaling","level":"3.1","depth":1,"next":{"title":"Feature Hashing","level":"3.2","depth":1,"path":"ft_engineering/hashing.md","ref":"ft_engineering/hashing.md","articles":[]},"previous":{"title":"English/Japanese Text Tokenizer","level":"2.3","depth":1,"path":"misc/tokenizer.md","ref":"misc/tokenizer.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","label":"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":"ft_engineering/scaling.md","mtime":"2016-10-22T16:43:44.000Z","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>
+