lychiyu/freqtrade_origin
1
0
Fork 0
mirror of https://github.com/freqtrade/freqtrade.git synced 2024-11-15 12:43:56 +00:00
Code Issues Actions Packages Projects Releases Wiki Activity
e0a538f380
freqtrade_origin/en/2022.9.1/freqai-feature-engineering/index.html

1577 lines
69 KiB
HTML
Raw Normal View History

Deployed 851d1e9da to 2022.9.1 in en with MkDocs 1.3.1 and mike 2.1.3
2024-08-20 18:04:06 +00:00
<!doctype html>
<html lang="en" class="no-js">
<head>
<meta charset="utf-8">
<meta name="viewport" content="width=device-width,initial-scale=1">
<link rel="canonical" href="https://www.freqtrade.io/2022.9.1/freqai-feature-engineering/">
<link rel="icon" href="../images/logo.png">
<meta name="generator" content="mkdocs-1.3.1, mkdocs-material-8.5.3">
<title>Feature engineering - Freqtrade</title>
<link rel="stylesheet" href="../assets/stylesheets/main.7a952b86.min.css">
<link rel="stylesheet" href="../assets/stylesheets/palette.cbb835fc.min.css">
<link rel="preconnect" href="https://fonts.gstatic.com" crossorigin>
<link rel="stylesheet" href="https://fonts.googleapis.com/css?family=Roboto:300,300i,400,400i,700,700i%7CRoboto+Mono:400,400i,700,700i&display=fallback">
<style>:root{--md-text-font:"Roboto";--md-code-font:"Roboto Mono"}</style>
<link rel="stylesheet" href="../stylesheets/ft.extra.css">
<script>__md_scope=new URL("..",location),__md_hash=e=>[...e].reduce((e,_)=>(e<<5)-e+_.charCodeAt(0),0),__md_get=(e,_=localStorage,t=__md_scope)=>JSON.parse(_.getItem(t.pathname+"."+e)),__md_set=(e,_,t=localStorage,a=__md_scope)=>{try{t.setItem(a.pathname+"."+e,JSON.stringify(_))}catch(e){}}</script>
</head>
<body dir="ltr" data-md-color-scheme="default" data-md-color-primary="blue-grey" data-md-color-accent="tear">
<script>var palette=__md_get("__palette");if(palette&&"object"==typeof palette.color)for(var key of Object.keys(palette.color))document.body.setAttribute("data-md-color-"+key,palette.color[key])</script>
<input class="md-toggle" data-md-toggle="drawer" type="checkbox" id="__drawer" autocomplete="off">
<input class="md-toggle" data-md-toggle="search" type="checkbox" id="__search" autocomplete="off">
<label class="md-overlay" for="__drawer"></label>
<div data-md-component="skip">
<a href="#feature-engineering" class="md-skip">
Skip to content
</a>
</div>
<div data-md-component="announce">
</div>
<div data-md-component="outdated" hidden>
</div>
<header class="md-header" data-md-component="header">
<nav class="md-header__inner md-grid" aria-label="Header">
<a href=".." title="Freqtrade" class="md-header__button md-logo" aria-label="Freqtrade" data-md-component="logo">
<img src="../images/logo.png" alt="logo">
</a>
<label class="md-header__button md-icon" for="__drawer">
<svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 24 24"><path d="M3 6h18v2H3V6m0 5h18v2H3v-2m0 5h18v2H3v-2Z"/></svg>
</label>
<div class="md-header__title" data-md-component="header-title">
<div class="md-header__ellipsis">
<div class="md-header__topic">
<span class="md-ellipsis">
Freqtrade
</span>
</div>
<div class="md-header__topic" data-md-component="header-topic">
<span class="md-ellipsis">
Feature engineering
</span>
</div>
</div>
</div>
<form class="md-header__option" data-md-component="palette">
<input class="md-option" data-md-color-media="" data-md-color-scheme="default" data-md-color-primary="blue-grey" data-md-color-accent="tear" aria-label="Switch to dark mode" type="radio" name="__palette" id="__palette_1">
<label class="md-header__button md-icon" title="Switch to dark mode" for="__palette_2" hidden>
<svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 24 24"><path d="M17 6H7c-3.31 0-6 2.69-6 6s2.69 6 6 6h10c3.31 0 6-2.69 6-6s-2.69-6-6-6zm0 10H7c-2.21 0-4-1.79-4-4s1.79-4 4-4h10c2.21 0 4 1.79 4 4s-1.79 4-4 4zM7 9c-1.66 0-3 1.34-3 3s1.34 3 3 3 3-1.34 3-3-1.34-3-3-3z"/></svg>
</label>
<input class="md-option" data-md-color-media="" data-md-color-scheme="slate" data-md-color-primary="blue-grey" data-md-color-accent="tear" aria-label="Switch to light mode" type="radio" name="__palette" id="__palette_2">
<label class="md-header__button md-icon" title="Switch to light mode" for="__palette_1" hidden>
<svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 24 24"><path d="M17 7H7a5 5 0 0 0-5 5 5 5 0 0 0 5 5h10a5 5 0 0 0 5-5 5 5 0 0 0-5-5m0 8a3 3 0 0 1-3-3 3 3 0 0 1 3-3 3 3 0 0 1 3 3 3 3 0 0 1-3 3Z"/></svg>
</label>
</form>
<label class="md-header__button md-icon" for="__search">
<svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 24 24"><path d="M9.5 3A6.5 6.5 0 0 1 16 9.5c0 1.61-.59 3.09-1.56 4.23l.27.27h.79l5 5-1.5 1.5-5-5v-.79l-.27-.27A6.516 6.516 0 0 1 9.5 16 6.5 6.5 0 0 1 3 9.5 6.5 6.5 0 0 1 9.5 3m0 2C7 5 5 7 5 9.5S7 14 9.5 14 14 12 14 9.5 12 5 9.5 5Z"/></svg>
</label>
<div class="md-search" data-md-component="search" role="dialog">
<label class="md-search__overlay" for="__search"></label>
<div class="md-search__inner" role="search">
<form class="md-search__form" name="search">
<input type="text" class="md-search__input" name="query" aria-label="Search" placeholder="Search" autocapitalize="off" autocorrect="off" autocomplete="off" spellcheck="false" data-md-component="search-query" required>
<label class="md-search__icon md-icon" for="__search">
<svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 24 24"><path d="M9.5 3A6.5 6.5 0 0 1 16 9.5c0 1.61-.59 3.09-1.56 4.23l.27.27h.79l5 5-1.5 1.5-5-5v-.79l-.27-.27A6.516 6.516 0 0 1 9.5 16 6.5 6.5 0 0 1 3 9.5 6.5 6.5 0 0 1 9.5 3m0 2C7 5 5 7 5 9.5S7 14 9.5 14 14 12 14 9.5 12 5 9.5 5Z"/></svg>
<svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 24 24"><path d="M20 11v2H8l5.5 5.5-1.42 1.42L4.16 12l7.92-7.92L13.5 5.5 8 11h12Z"/></svg>
</label>
<nav class="md-search__options" aria-label="Search">
<a href="javascript:void(0)" class="md-search__icon md-icon" title="Share" aria-label="Share" data-clipboard data-clipboard-text="" data-md-component="search-share" tabindex="-1">
<svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 24 24"><path d="M18 16.08c-.76 0-1.44.3-1.96.77L8.91 12.7c.05-.23.09-.46.09-.7 0-.24-.04-.47-.09-.7l7.05-4.11c.54.5 1.25.81 2.04.81a3 3 0 0 0 3-3 3 3 0 0 0-3-3 3 3 0 0 0-3 3c0 .24.04.47.09.7L8.04 9.81C7.5 9.31 6.79 9 6 9a3 3 0 0 0-3 3 3 3 0 0 0 3 3c.79 0 1.5-.31 2.04-.81l7.12 4.15c-.05.21-.08.43-.08.66 0 1.61 1.31 2.91 2.92 2.91 1.61 0 2.92-1.3 2.92-2.91A2.92 2.92 0 0 0 18 16.08Z"/></svg>
</a>
<button type="reset" class="md-search__icon md-icon" title="Clear" aria-label="Clear" tabindex="-1">
<svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 24 24"><path d="M19 6.41 17.59 5 12 10.59 6.41 5 5 6.41 10.59 12 5 17.59 6.41 19 12 13.41 17.59 19 19 17.59 13.41 12 19 6.41Z"/></svg>
</button>
</nav>
</form>
<div class="md-search__output">
<div class="md-search__scrollwrap" data-md-scrollfix>
<div class="md-search-result" data-md-component="search-result">
<div class="md-search-result__meta">
Initializing search
</div>
<ol class="md-search-result__list"></ol>
</div>
</div>
</div>
</div>
</div>
<div class="md-header__source">
<a href="https://github.com/freqtrade/freqtrade" title="Go to repository" class="md-source" data-md-component="source">
<div class="md-source__icon md-icon">
<svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 448 512"><!--! Font Awesome Free 6.2.0 by @fontawesome - https://fontawesome.com License - https://fontawesome.com/license/free (Icons: CC BY 4.0, Fonts: SIL OFL 1.1, Code: MIT License) Copyright 2022 Fonticons, Inc.--><path d="M439.55 236.05 244 40.45a28.87 28.87 0 0 0-40.81 0l-40.66 40.63 51.52 51.52c27.06-9.14 52.68 16.77 43.39 43.68l49.66 49.66c34.23-11.8 61.18 31 35.47 56.69-26.49 26.49-70.21-2.87-56-37.34L240.22 199v121.85c25.3 12.54 22.26 41.85 9.08 55a34.34 34.34 0 0 1-48.55 0c-17.57-17.6-11.07-46.91 11.25-56v-123c-20.8-8.51-24.6-30.74-18.64-45L142.57 101 8.45 235.14a28.86 28.86 0 0 0 0 40.81l195.61 195.6a28.86 28.86 0 0 0 40.8 0l194.69-194.69a28.86 28.86 0 0 0 0-40.81z"/></svg>
</div>
<div class="md-source__repository">
GitHub
</div>
</a>
</div>
</nav>
</header>
<div class="md-container" data-md-component="container">
<main class="md-main" data-md-component="main">
<div class="md-main__inner md-grid">
<!-- Main navigation -->
<div class="md-sidebar md-sidebar--primary" data-md-component="sidebar" data-md-type="navigation" >
<div class="md-sidebar__scrollwrap">
<div id="widget-wrapper">
</div>
<div class="md-sidebar__inner">
<nav class="md-nav md-nav--primary" aria-label="Navigation" data-md-level="0">
<label class="md-nav__title" for="__drawer">
<a href=".." title="Freqtrade" class="md-nav__button md-logo" aria-label="Freqtrade" data-md-component="logo">
<img src="../images/logo.png" alt="logo">
</a>
Freqtrade
</label>
<div class="md-nav__source">
<a href="https://github.com/freqtrade/freqtrade" title="Go to repository" class="md-source" data-md-component="source">
<div class="md-source__icon md-icon">
<svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 448 512"><!--! Font Awesome Free 6.2.0 by @fontawesome - https://fontawesome.com License - https://fontawesome.com/license/free (Icons: CC BY 4.0, Fonts: SIL OFL 1.1, Code: MIT License) Copyright 2022 Fonticons, Inc.--><path d="M439.55 236.05 244 40.45a28.87 28.87 0 0 0-40.81 0l-40.66 40.63 51.52 51.52c27.06-9.14 52.68 16.77 43.39 43.68l49.66 49.66c34.23-11.8 61.18 31 35.47 56.69-26.49 26.49-70.21-2.87-56-37.34L240.22 199v121.85c25.3 12.54 22.26 41.85 9.08 55a34.34 34.34 0 0 1-48.55 0c-17.57-17.6-11.07-46.91 11.25-56v-123c-20.8-8.51-24.6-30.74-18.64-45L142.57 101 8.45 235.14a28.86 28.86 0 0 0 0 40.81l195.61 195.6a28.86 28.86 0 0 0 40.8 0l194.69-194.69a28.86 28.86 0 0 0 0-40.81z"/></svg>
</div>
<div class="md-source__repository">
GitHub
</div>
</a>
</div>
<ul class="md-nav__list" data-md-scrollfix>
<li class="md-nav__item">
<a href=".." class="md-nav__link">
Home
</a>
</li>
<li class="md-nav__item">
<a href="../docker_quickstart/" class="md-nav__link">
Quickstart with Docker
</a>
</li>
<li class="md-nav__item md-nav__item--nested">
<input class="md-nav__toggle md-toggle" data-md-toggle="__nav_3" type="checkbox" id="__nav_3" >
<label class="md-nav__link" for="__nav_3">
Installation
<span class="md-nav__icon md-icon"></span>
</label>
<nav class="md-nav" aria-label="Installation" data-md-level="1">
<label class="md-nav__title" for="__nav_3">
<span class="md-nav__icon md-icon"></span>
Installation
</label>
<ul class="md-nav__list" data-md-scrollfix>
<li class="md-nav__item">
<a href="../installation/" class="md-nav__link">
Linux/MacOS/Raspberry
</a>
</li>
<li class="md-nav__item">
<a href="../windows_installation/" class="md-nav__link">
Windows
</a>
</li>
</ul>
</nav>
</li>
<li class="md-nav__item">
<a href="../bot-basics/" class="md-nav__link">
Freqtrade Basics
</a>
</li>
<li class="md-nav__item">
<a href="../configuration/" class="md-nav__link">
Configuration
</a>
</li>
<li class="md-nav__item">
<a href="../strategy-customization/" class="md-nav__link">
Strategy Customization
</a>
</li>
<li class="md-nav__item">
<a href="../strategy-callbacks/" class="md-nav__link">
Strategy Callbacks
</a>
</li>
<li class="md-nav__item">
<a href="../stoploss/" class="md-nav__link">
Stoploss
</a>
</li>
<li class="md-nav__item">
<a href="../plugins/" class="md-nav__link">
Plugins
</a>
</li>
<li class="md-nav__item">
<a href="../bot-usage/" class="md-nav__link">
Start the bot
</a>
</li>
<li class="md-nav__item md-nav__item--nested">
<input class="md-nav__toggle md-toggle" data-md-toggle="__nav_11" type="checkbox" id="__nav_11" >
<label class="md-nav__link" for="__nav_11">
Control the bot
<span class="md-nav__icon md-icon"></span>
</label>
<nav class="md-nav" aria-label="Control the bot" data-md-level="1">
<label class="md-nav__title" for="__nav_11">
<span class="md-nav__icon md-icon"></span>
Control the bot
</label>
<ul class="md-nav__list" data-md-scrollfix>
<li class="md-nav__item">
<a href="../telegram-usage/" class="md-nav__link">
Telegram
</a>
</li>
<li class="md-nav__item">
<a href="../rest-api/" class="md-nav__link">
REST API & FreqUI
</a>
</li>
<li class="md-nav__item">
<a href="../webhook-config/" class="md-nav__link">
Web Hook
</a>
</li>
</ul>
</nav>
</li>
<li class="md-nav__item">
<a href="../data-download/" class="md-nav__link">
Data Downloading
</a>
</li>
<li class="md-nav__item">
<a href="../backtesting/" class="md-nav__link">
Backtesting
</a>
</li>
<li class="md-nav__item">
<a href="../hyperopt/" class="md-nav__link">
Hyperopt
</a>
</li>
<li class="md-nav__item md-nav__item--active md-nav__item--nested">
<input class="md-nav__toggle md-toggle" data-md-toggle="__nav_15" type="checkbox" id="__nav_15" checked>
<label class="md-nav__link" for="__nav_15">
FreqAI
<span class="md-nav__icon md-icon"></span>
</label>
<nav class="md-nav" aria-label="FreqAI" data-md-level="1">
<label class="md-nav__title" for="__nav_15">
<span class="md-nav__icon md-icon"></span>
FreqAI
</label>
<ul class="md-nav__list" data-md-scrollfix>
<li class="md-nav__item">
<a href="../freqai/" class="md-nav__link">
Introduction
</a>
</li>
<li class="md-nav__item">
<a href="../freqai-configuration/" class="md-nav__link">
Configuration
</a>
</li>
<li class="md-nav__item">
<a href="../freqai-parameter-table/" class="md-nav__link">
Parameter table
</a>
</li>
<li class="md-nav__item md-nav__item--active">
<input class="md-nav__toggle md-toggle" data-md-toggle="toc" type="checkbox" id="__toc">
<label class="md-nav__link md-nav__link--active" for="__toc">
Feature engineering
<span class="md-nav__icon md-icon"></span>
</label>
<a href="./" class="md-nav__link md-nav__link--active">
Feature engineering
</a>
<nav class="md-nav md-nav--secondary" aria-label="Table of contents">
<label class="md-nav__title" for="__toc">
<span class="md-nav__icon md-icon"></span>
Table of contents
</label>
<ul class="md-nav__list" data-md-component="toc" data-md-scrollfix>
<li class="md-nav__item">
<a href="#defining-the-features" class="md-nav__link">
Defining the features
</a>
<nav class="md-nav" aria-label="Defining the features">
<ul class="md-nav__list">
<li class="md-nav__item">
<a href="#returning-additional-info-from-training" class="md-nav__link">
Returning additional info from training
</a>
</li>
</ul>
</nav>
</li>
<li class="md-nav__item">
<a href="#feature-normalization" class="md-nav__link">
Feature normalization
</a>
</li>
<li class="md-nav__item">
<a href="#data-dimensionality-reduction-with-principal-component-analysis" class="md-nav__link">
Data dimensionality reduction with Principal Component Analysis
</a>
</li>
<li class="md-nav__item">
<a href="#inlier-metric" class="md-nav__link">
Inlier metric
</a>
</li>
<li class="md-nav__item">
<a href="#weighting-features-for-temporal-importance" class="md-nav__link">
Weighting features for temporal importance
</a>
</li>
<li class="md-nav__item">
<a href="#outlier-detection" class="md-nav__link">
Outlier detection
</a>
<nav class="md-nav" aria-label="Outlier detection">
<ul class="md-nav__list">
<li class="md-nav__item">
<a href="#identifying-outliers-with-the-dissimilarity-index-di" class="md-nav__link">
Identifying outliers with the Dissimilarity Index (DI)
</a>
</li>
<li class="md-nav__item">
<a href="#identifying-outliers-using-a-support-vector-machine-svm" class="md-nav__link">
Identifying outliers using a Support Vector Machine (SVM)
</a>
</li>
<li class="md-nav__item">
<a href="#identifying-outliers-with-dbscan" class="md-nav__link">
Identifying outliers with DBSCAN
</a>
</li>
</ul>
</nav>
</li>
</ul>
</nav>
</li>
<li class="md-nav__item">
<a href="../freqai-running/" class="md-nav__link">
Running FreqAI
</a>
</li>
<li class="md-nav__item">
<a href="../freqai-developers/" class="md-nav__link">
Developer guide
</a>
</li>
</ul>
</nav>
</li>
<li class="md-nav__item">
<a href="../leverage/" class="md-nav__link">
Short / Leverage
</a>
</li>
<li class="md-nav__item">
<a href="../utils/" class="md-nav__link">
Utility Sub-commands
</a>
</li>
<li class="md-nav__item">
<a href="../plotting/" class="md-nav__link">
Plotting
</a>
</li>
<li class="md-nav__item">
<a href="../exchanges/" class="md-nav__link">
Exchange-specific Notes
</a>
</li>
<li class="md-nav__item md-nav__item--nested">
<input class="md-nav__toggle md-toggle" data-md-toggle="__nav_20" type="checkbox" id="__nav_20" >
<label class="md-nav__link" for="__nav_20">
Data Analysis
<span class="md-nav__icon md-icon"></span>
</label>
<nav class="md-nav" aria-label="Data Analysis" data-md-level="1">
<label class="md-nav__title" for="__nav_20">
<span class="md-nav__icon md-icon"></span>
Data Analysis
</label>
<ul class="md-nav__list" data-md-scrollfix>
<li class="md-nav__item">
<a href="../data-analysis/" class="md-nav__link">
Jupyter Notebooks
</a>
</li>
<li class="md-nav__item">
<a href="../strategy_analysis_example/" class="md-nav__link">
Strategy analysis
</a>
</li>
<li class="md-nav__item">
<a href="../advanced-backtesting/" class="md-nav__link">
Backtest analysis
</a>
</li>
</ul>
</nav>
</li>
<li class="md-nav__item md-nav__item--nested">
<input class="md-nav__toggle md-toggle" data-md-toggle="__nav_21" type="checkbox" id="__nav_21" >
<label class="md-nav__link" for="__nav_21">
Advanced Topics
<span class="md-nav__icon md-icon"></span>
</label>
<nav class="md-nav" aria-label="Advanced Topics" data-md-level="1">
<label class="md-nav__title" for="__nav_21">
<span class="md-nav__icon md-icon"></span>
Advanced Topics
</label>
<ul class="md-nav__list" data-md-scrollfix>
<li class="md-nav__item">
<a href="../advanced-setup/" class="md-nav__link">
Advanced Post-installation Tasks
</a>
</li>
<li class="md-nav__item">
<a href="../strategy-advanced/" class="md-nav__link">
Advanced Strategy
</a>
</li>
<li class="md-nav__item">
<a href="../advanced-hyperopt/" class="md-nav__link">
Advanced Hyperopt
</a>
</li>
<li class="md-nav__item">
<a href="../producer-consumer/" class="md-nav__link">
Producer/Consumer mode
</a>
</li>
<li class="md-nav__item">
<a href="../edge/" class="md-nav__link">
Edge Positioning
</a>
</li>
<li class="md-nav__item">
<a href="../sandbox-testing/" class="md-nav__link">
Sandbox Testing
</a>
</li>
</ul>
</nav>
</li>
<li class="md-nav__item">
<a href="../faq/" class="md-nav__link">
FAQ
</a>
</li>
<li class="md-nav__item">
<a href="../sql_cheatsheet/" class="md-nav__link">
SQL Cheat-sheet
</a>
</li>
<li class="md-nav__item">
<a href="../strategy_migration/" class="md-nav__link">
Strategy migration
</a>
</li>
<li class="md-nav__item">
<a href="../updating/" class="md-nav__link">
Updating Freqtrade
</a>
</li>
<li class="md-nav__item">
<a href="../deprecated/" class="md-nav__link">
Deprecated Features
</a>
</li>
<li class="md-nav__item">
<a href="../developer/" class="md-nav__link">
Contributors Guide
</a>
</li>
</ul>
</nav>
</div>
</div>
</div>
<!-- Table of contents -->
<div class="md-sidebar md-sidebar--secondary" data-md-component="sidebar" data-md-type="toc" >
<div class="md-sidebar__scrollwrap">
<div class="md-sidebar__inner">
<nav class="md-nav md-nav--secondary" aria-label="Table of contents">
<label class="md-nav__title" for="__toc">
<span class="md-nav__icon md-icon"></span>
Table of contents
</label>
<ul class="md-nav__list" data-md-component="toc" data-md-scrollfix>
<li class="md-nav__item">
<a href="#defining-the-features" class="md-nav__link">
Defining the features
</a>
<nav class="md-nav" aria-label="Defining the features">
<ul class="md-nav__list">
<li class="md-nav__item">
<a href="#returning-additional-info-from-training" class="md-nav__link">
Returning additional info from training
</a>
</li>
</ul>
</nav>
</li>
<li class="md-nav__item">
<a href="#feature-normalization" class="md-nav__link">
Feature normalization
</a>
</li>
<li class="md-nav__item">
<a href="#data-dimensionality-reduction-with-principal-component-analysis" class="md-nav__link">
Data dimensionality reduction with Principal Component Analysis
</a>
</li>
<li class="md-nav__item">
<a href="#inlier-metric" class="md-nav__link">
Inlier metric
</a>
</li>
<li class="md-nav__item">
<a href="#weighting-features-for-temporal-importance" class="md-nav__link">
Weighting features for temporal importance
</a>
</li>
<li class="md-nav__item">
<a href="#outlier-detection" class="md-nav__link">
Outlier detection
</a>
<nav class="md-nav" aria-label="Outlier detection">
<ul class="md-nav__list">
<li class="md-nav__item">
<a href="#identifying-outliers-with-the-dissimilarity-index-di" class="md-nav__link">
Identifying outliers with the Dissimilarity Index (DI)
</a>
</li>
<li class="md-nav__item">
<a href="#identifying-outliers-using-a-support-vector-machine-svm" class="md-nav__link">
Identifying outliers using a Support Vector Machine (SVM)
</a>
</li>
<li class="md-nav__item">
<a href="#identifying-outliers-with-dbscan" class="md-nav__link">
Identifying outliers with DBSCAN
</a>
</li>
</ul>
</nav>
</li>
</ul>
</nav>
</div>
</div>
</div>
<div class="md-content" data-md-component="content">
<article class="md-content__inner md-typeset">
<a href="https://github.com/freqtrade/freqtrade/edit/develop/docs/freqai-feature-engineering.md" title="Edit this page" class="md-content__button md-icon">
<svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 24 24"><path d="M20.71 7.04c.39-.39.39-1.04 0-1.41l-2.34-2.34c-.37-.39-1.02-.39-1.41 0l-1.84 1.83 3.75 3.75M3 17.25V21h3.75L17.81 9.93l-3.75-3.75L3 17.25Z"/></svg>
</a>
<h1 id="feature-engineering">Feature engineering<a class="headerlink" href="#feature-engineering" title="Permanent link">&para;</a></h1>
<h2 id="defining-the-features">Defining the features<a class="headerlink" href="#defining-the-features" title="Permanent link">&para;</a></h2>
<p>Low level feature engineering is performed in the user strategy within a function called <code>populate_any_indicators()</code>. That function sets the <code>base features</code> such as, <code>RSI</code>, <code>MFI</code>, <code>EMA</code>, <code>SMA</code>, time of day, volume, etc. The <code>base features</code> can be custom indicators or they can be imported from any technical-analysis library that you can find. One important syntax rule is that all <code>base features</code> string names are prepended with <code>%</code>, while labels/targets are prepended with <code>&amp;</code>.</p>
<p>Meanwhile, high level feature engineering is handled within <code>"feature_parameters":{}</code> in the <code>FreqAI</code> config. Within this file, it is possible to decide large scale feature expansions on top of the <code>base_features</code> such as "including correlated pairs" or "including informative timeframes" or even "including recent candles." </p>
<p>It is advisable to start from the template <code>populate_any_indicators()</code> in the source provided example strategy (found in <code>templates/FreqaiExampleStrategy.py</code>) to ensure that the feature definitions are following the correct conventions. Here is an example of how to set the indicators and labels in the strategy:</p>
<div class="highlight"><pre><span></span><code> <span class="k">def</span> <span class="nf">populate_any_indicators</span><span class="p">(</span>
<span class="bp">self</span><span class="p">,</span> <span class="n">pair</span><span class="p">,</span> <span class="n">df</span><span class="p">,</span> <span class="n">tf</span><span class="p">,</span> <span class="n">informative</span><span class="o">=</span><span class="kc">None</span><span class="p">,</span> <span class="n">set_generalized_indicators</span><span class="o">=</span><span class="kc">False</span>
<span class="p">):</span>
<span class="w"> </span><span class="sd">&quot;&quot;&quot;</span>
<span class="sd"> Function designed to automatically generate, name, and merge features</span>
<span class="sd"> from user-indicated timeframes in the configuration file. The user controls the indicators</span>
<span class="sd"> passed to the training/prediction by prepending indicators with `&#39;%-&#39; + coin `</span>
<span class="sd"> (see convention below). I.e., the user should not prepend any supporting metrics</span>
<span class="sd"> (e.g., bb_lowerband below) with % unless they explicitly want to pass that metric to the</span>
<span class="sd"> model.</span>
<span class="sd"> :param pair: pair to be used as informative</span>
<span class="sd"> :param df: strategy dataframe which will receive merges from informatives</span>
<span class="sd"> :param tf: timeframe of the dataframe which will modify the feature names</span>
<span class="sd"> :param informative: the dataframe associated with the informative pair</span>
<span class="sd"> :param coin: the name of the coin which will modify the feature names.</span>
<span class="sd"> &quot;&quot;&quot;</span>
<span class="n">coin</span> <span class="o">=</span> <span class="n">pair</span><span class="o">.</span><span class="n">split</span><span class="p">(</span><span class="s1">&#39;/&#39;</span><span class="p">)[</span><span class="mi">0</span><span class="p">]</span>
<span class="k">if</span> <span class="n">informative</span> <span class="ow">is</span> <span class="kc">None</span><span class="p">:</span>
<span class="n">informative</span> <span class="o">=</span> <span class="bp">self</span><span class="o">.</span><span class="n">dp</span><span class="o">.</span><span class="n">get_pair_dataframe</span><span class="p">(</span><span class="n">pair</span><span class="p">,</span> <span class="n">tf</span><span class="p">)</span>
<span class="c1"># first loop is automatically duplicating indicators for time periods</span>
<span class="k">for</span> <span class="n">t</span> <span class="ow">in</span> <span class="bp">self</span><span class="o">.</span><span class="n">freqai_info</span><span class="p">[</span><span class="s2">&quot;feature_parameters&quot;</span><span class="p">][</span><span class="s2">&quot;indicator_periods_candles&quot;</span><span class="p">]:</span>
<span class="n">t</span> <span class="o">=</span> <span class="nb">int</span><span class="p">(</span><span class="n">t</span><span class="p">)</span>
<span class="n">informative</span><span class="p">[</span><span class="sa">f</span><span class="s2">&quot;%-</span><span class="si">{</span><span class="n">coin</span><span class="si">}</span><span class="s2">rsi-period_</span><span class="si">{</span><span class="n">t</span><span class="si">}</span><span class="s2">&quot;</span><span class="p">]</span> <span class="o">=</span> <span class="n">ta</span><span class="o">.</span><span class="n">RSI</span><span class="p">(</span><span class="n">informative</span><span class="p">,</span> <span class="n">timeperiod</span><span class="o">=</span><span class="n">t</span><span class="p">)</span>
<span class="n">informative</span><span class="p">[</span><span class="sa">f</span><span class="s2">&quot;%-</span><span class="si">{</span><span class="n">coin</span><span class="si">}</span><span class="s2">mfi-period_</span><span class="si">{</span><span class="n">t</span><span class="si">}</span><span class="s2">&quot;</span><span class="p">]</span> <span class="o">=</span> <span class="n">ta</span><span class="o">.</span><span class="n">MFI</span><span class="p">(</span><span class="n">informative</span><span class="p">,</span> <span class="n">timeperiod</span><span class="o">=</span><span class="n">t</span><span class="p">)</span>
<span class="n">informative</span><span class="p">[</span><span class="sa">f</span><span class="s2">&quot;%-</span><span class="si">{</span><span class="n">coin</span><span class="si">}</span><span class="s2">adx-period_</span><span class="si">{</span><span class="n">t</span><span class="si">}</span><span class="s2">&quot;</span><span class="p">]</span> <span class="o">=</span> <span class="n">ta</span><span class="o">.</span><span class="n">ADX</span><span class="p">(</span><span class="n">informative</span><span class="p">,</span> <span class="n">window</span><span class="o">=</span><span class="n">t</span><span class="p">)</span>
<span class="n">bollinger</span> <span class="o">=</span> <span class="n">qtpylib</span><span class="o">.</span><span class="n">bollinger_bands</span><span class="p">(</span>
<span class="n">qtpylib</span><span class="o">.</span><span class="n">typical_price</span><span class="p">(</span><span class="n">informative</span><span class="p">),</span> <span class="n">window</span><span class="o">=</span><span class="n">t</span><span class="p">,</span> <span class="n">stds</span><span class="o">=</span><span class="mf">2.2</span>
<span class="p">)</span>
<span class="n">informative</span><span class="p">[</span><span class="sa">f</span><span class="s2">&quot;</span><span class="si">{</span><span class="n">coin</span><span class="si">}</span><span class="s2">bb_lowerband-period_</span><span class="si">{</span><span class="n">t</span><span class="si">}</span><span class="s2">&quot;</span><span class="p">]</span> <span class="o">=</span> <span class="n">bollinger</span><span class="p">[</span><span class="s2">&quot;lower&quot;</span><span class="p">]</span>
<span class="n">informative</span><span class="p">[</span><span class="sa">f</span><span class="s2">&quot;</span><span class="si">{</span><span class="n">coin</span><span class="si">}</span><span class="s2">bb_middleband-period_</span><span class="si">{</span><span class="n">t</span><span class="si">}</span><span class="s2">&quot;</span><span class="p">]</span> <span class="o">=</span> <span class="n">bollinger</span><span class="p">[</span><span class="s2">&quot;mid&quot;</span><span class="p">]</span>
<span class="n">informative</span><span class="p">[</span><span class="sa">f</span><span class="s2">&quot;</span><span class="si">{</span><span class="n">coin</span><span class="si">}</span><span class="s2">bb_upperband-period_</span><span class="si">{</span><span class="n">t</span><span class="si">}</span><span class="s2">&quot;</span><span class="p">]</span> <span class="o">=</span> <span class="n">bollinger</span><span class="p">[</span><span class="s2">&quot;upper&quot;</span><span class="p">]</span>
<span class="n">informative</span><span class="p">[</span><span class="sa">f</span><span class="s2">&quot;%-</span><span class="si">{</span><span class="n">coin</span><span class="si">}</span><span class="s2">bb_width-period_</span><span class="si">{</span><span class="n">t</span><span class="si">}</span><span class="s2">&quot;</span><span class="p">]</span> <span class="o">=</span> <span class="p">(</span>
<span class="n">informative</span><span class="p">[</span><span class="sa">f</span><span class="s2">&quot;</span><span class="si">{</span><span class="n">coin</span><span class="si">}</span><span class="s2">bb_upperband-period_</span><span class="si">{</span><span class="n">t</span><span class="si">}</span><span class="s2">&quot;</span><span class="p">]</span>
<span class="o">-</span> <span class="n">informative</span><span class="p">[</span><span class="sa">f</span><span class="s2">&quot;</span><span class="si">{</span><span class="n">coin</span><span class="si">}</span><span class="s2">bb_lowerband-period_</span><span class="si">{</span><span class="n">t</span><span class="si">}</span><span class="s2">&quot;</span><span class="p">]</span>
<span class="p">)</span> <span class="o">/</span> <span class="n">informative</span><span class="p">[</span><span class="sa">f</span><span class="s2">&quot;</span><span class="si">{</span><span class="n">coin</span><span class="si">}</span><span class="s2">bb_middleband-period_</span><span class="si">{</span><span class="n">t</span><span class="si">}</span><span class="s2">&quot;</span><span class="p">]</span>
<span class="n">informative</span><span class="p">[</span><span class="sa">f</span><span class="s2">&quot;%-</span><span class="si">{</span><span class="n">coin</span><span class="si">}</span><span class="s2">close-bb_lower-period_</span><span class="si">{</span><span class="n">t</span><span class="si">}</span><span class="s2">&quot;</span><span class="p">]</span> <span class="o">=</span> <span class="p">(</span>
<span class="n">informative</span><span class="p">[</span><span class="s2">&quot;close&quot;</span><span class="p">]</span> <span class="o">/</span> <span class="n">informative</span><span class="p">[</span><span class="sa">f</span><span class="s2">&quot;</span><span class="si">{</span><span class="n">coin</span><span class="si">}</span><span class="s2">bb_lowerband-period_</span><span class="si">{</span><span class="n">t</span><span class="si">}</span><span class="s2">&quot;</span><span class="p">]</span>
<span class="p">)</span>
<span class="n">informative</span><span class="p">[</span><span class="sa">f</span><span class="s2">&quot;%-</span><span class="si">{</span><span class="n">coin</span><span class="si">}</span><span class="s2">relative_volume-period_</span><span class="si">{</span><span class="n">t</span><span class="si">}</span><span class="s2">&quot;</span><span class="p">]</span> <span class="o">=</span> <span class="p">(</span>
<span class="n">informative</span><span class="p">[</span><span class="s2">&quot;volume&quot;</span><span class="p">]</span> <span class="o">/</span> <span class="n">informative</span><span class="p">[</span><span class="s2">&quot;volume&quot;</span><span class="p">]</span><span class="o">.</span><span class="n">rolling</span><span class="p">(</span><span class="n">t</span><span class="p">)</span><span class="o">.</span><span class="n">mean</span><span class="p">()</span>
<span class="p">)</span>
<span class="n">indicators</span> <span class="o">=</span> <span class="p">[</span><span class="n">col</span> <span class="k">for</span> <span class="n">col</span> <span class="ow">in</span> <span class="n">informative</span> <span class="k">if</span> <span class="n">col</span><span class="o">.</span><span class="n">startswith</span><span class="p">(</span><span class="s2">&quot;%&quot;</span><span class="p">)]</span>
<span class="c1"># This loop duplicates and shifts all indicators to add a sense of recency to data</span>
<span class="k">for</span> <span class="n">n</span> <span class="ow">in</span> <span class="nb">range</span><span class="p">(</span><span class="bp">self</span><span class="o">.</span><span class="n">freqai_info</span><span class="p">[</span><span class="s2">&quot;feature_parameters&quot;</span><span class="p">][</span><span class="s2">&quot;include_shifted_candles&quot;</span><span class="p">]</span> <span class="o">+</span> <span class="mi">1</span><span class="p">):</span>
<span class="k">if</span> <span class="n">n</span> <span class="o">==</span> <span class="mi">0</span><span class="p">:</span>
<span class="k">continue</span>
<span class="n">informative_shift</span> <span class="o">=</span> <span class="n">informative</span><span class="p">[</span><span class="n">indicators</span><span class="p">]</span><span class="o">.</span><span class="n">shift</span><span class="p">(</span><span class="n">n</span><span class="p">)</span>
<span class="n">informative_shift</span> <span class="o">=</span> <span class="n">informative_shift</span><span class="o">.</span><span class="n">add_suffix</span><span class="p">(</span><span class="s2">&quot;_shift-&quot;</span> <span class="o">+</span> <span class="nb">str</span><span class="p">(</span><span class="n">n</span><span class="p">))</span>
<span class="n">informative</span> <span class="o">=</span> <span class="n">pd</span><span class="o">.</span><span class="n">concat</span><span class="p">((</span><span class="n">informative</span><span class="p">,</span> <span class="n">informative_shift</span><span class="p">),</span> <span class="n">axis</span><span class="o">=</span><span class="mi">1</span><span class="p">)</span>
<span class="n">df</span> <span class="o">=</span> <span class="n">merge_informative_pair</span><span class="p">(</span><span class="n">df</span><span class="p">,</span> <span class="n">informative</span><span class="p">,</span> <span class="bp">self</span><span class="o">.</span><span class="n">config</span><span class="p">[</span><span class="s2">&quot;timeframe&quot;</span><span class="p">],</span> <span class="n">tf</span><span class="p">,</span> <span class="n">ffill</span><span class="o">=</span><span class="kc">True</span><span class="p">)</span>
<span class="n">skip_columns</span> <span class="o">=</span> <span class="p">[</span>
<span class="p">(</span><span class="n">s</span> <span class="o">+</span> <span class="s2">&quot;_&quot;</span> <span class="o">+</span> <span class="n">tf</span><span class="p">)</span> <span class="k">for</span> <span class="n">s</span> <span class="ow">in</span> <span class="p">[</span><span class="s2">&quot;date&quot;</span><span class="p">,</span> <span class="s2">&quot;open&quot;</span><span class="p">,</span> <span class="s2">&quot;high&quot;</span><span class="p">,</span> <span class="s2">&quot;low&quot;</span><span class="p">,</span> <span class="s2">&quot;close&quot;</span><span class="p">,</span> <span class="s2">&quot;volume&quot;</span><span class="p">]</span>
<span class="p">]</span>
<span class="n">df</span> <span class="o">=</span> <span class="n">df</span><span class="o">.</span><span class="n">drop</span><span class="p">(</span><span class="n">columns</span><span class="o">=</span><span class="n">skip_columns</span><span class="p">)</span>
<span class="c1"># Add generalized indicators here (because in live, it will call this</span>
<span class="c1"># function to populate indicators during training). Notice how we ensure not to</span>
<span class="c1"># add them multiple times</span>
<span class="k">if</span> <span class="n">set_generalized_indicators</span><span class="p">:</span>
<span class="n">df</span><span class="p">[</span><span class="s2">&quot;</span><span class="si">%-d</span><span class="s2">ay_of_week&quot;</span><span class="p">]</span> <span class="o">=</span> <span class="p">(</span><span class="n">df</span><span class="p">[</span><span class="s2">&quot;date&quot;</span><span class="p">]</span><span class="o">.</span><span class="n">dt</span><span class="o">.</span><span class="n">dayofweek</span> <span class="o">+</span> <span class="mi">1</span><span class="p">)</span> <span class="o">/</span> <span class="mi">7</span>
<span class="n">df</span><span class="p">[</span><span class="s2">&quot;</span><span class="si">%-ho</span><span class="s2">ur_of_day&quot;</span><span class="p">]</span> <span class="o">=</span> <span class="p">(</span><span class="n">df</span><span class="p">[</span><span class="s2">&quot;date&quot;</span><span class="p">]</span><span class="o">.</span><span class="n">dt</span><span class="o">.</span><span class="n">hour</span> <span class="o">+</span> <span class="mi">1</span><span class="p">)</span> <span class="o">/</span> <span class="mi">25</span>
<span class="c1"># user adds targets here by prepending them with &amp;- (see convention below)</span>
<span class="c1"># If user wishes to use multiple targets, a multioutput prediction model</span>
<span class="c1"># needs to be used such as templates/CatboostPredictionMultiModel.py</span>
<span class="n">df</span><span class="p">[</span><span class="s2">&quot;&amp;-s_close&quot;</span><span class="p">]</span> <span class="o">=</span> <span class="p">(</span>
<span class="n">df</span><span class="p">[</span><span class="s2">&quot;close&quot;</span><span class="p">]</span>
<span class="o">.</span><span class="n">shift</span><span class="p">(</span><span class="o">-</span><span class="bp">self</span><span class="o">.</span><span class="n">freqai_info</span><span class="p">[</span><span class="s2">&quot;feature_parameters&quot;</span><span class="p">][</span><span class="s2">&quot;label_period_candles&quot;</span><span class="p">])</span>
<span class="o">.</span><span class="n">rolling</span><span class="p">(</span><span class="bp">self</span><span class="o">.</span><span class="n">freqai_info</span><span class="p">[</span><span class="s2">&quot;feature_parameters&quot;</span><span class="p">][</span><span class="s2">&quot;label_period_candles&quot;</span><span class="p">])</span>
<span class="o">.</span><span class="n">mean</span><span class="p">()</span>
<span class="o">/</span> <span class="n">df</span><span class="p">[</span><span class="s2">&quot;close&quot;</span><span class="p">]</span>
<span class="o">-</span> <span class="mi">1</span>
<span class="p">)</span>
<span class="k">return</span> <span class="n">df</span>
</code></pre></div>
<p>In the presented example, the user does not wish to pass the <code>bb_lowerband</code> as a feature to the model,
and has therefore not prepended it with <code>%</code>. The user does, however, wish to pass <code>bb_width</code> to the
model for training/prediction and has therefore prepended it with <code>%</code>.</p>
<p>After having defined the <code>base features</code>, the next step is to expand upon them using the powerful <code>feature_parameters</code> in the configuration file:</p>
<div class="highlight"><pre><span></span><code><span class="w"> </span><span class="nt">&quot;freqai&quot;</span><span class="p">:</span><span class="w"> </span><span class="p">{</span>
<span class="w"> </span><span class="c1">//...</span>
<span class="w"> </span><span class="nt">&quot;feature_parameters&quot;</span><span class="w"> </span><span class="p">:</span><span class="w"> </span><span class="p">{</span>
<span class="w"> </span><span class="nt">&quot;include_timeframes&quot;</span><span class="p">:</span><span class="w"> </span><span class="p">[</span><span class="s2">&quot;5m&quot;</span><span class="p">,</span><span class="s2">&quot;15m&quot;</span><span class="p">,</span><span class="s2">&quot;4h&quot;</span><span class="p">],</span>
<span class="w"> </span><span class="nt">&quot;include_corr_pairlist&quot;</span><span class="p">:</span><span class="w"> </span><span class="p">[</span>
<span class="w"> </span><span class="s2">&quot;ETH/USD&quot;</span><span class="p">,</span>
<span class="w"> </span><span class="s2">&quot;LINK/USD&quot;</span><span class="p">,</span>
<span class="w"> </span><span class="s2">&quot;BNB/USD&quot;</span>
<span class="w"> </span><span class="p">],</span>
<span class="w"> </span><span class="nt">&quot;label_period_candles&quot;</span><span class="p">:</span><span class="w"> </span><span class="mi">24</span><span class="p">,</span>
<span class="w"> </span><span class="nt">&quot;include_shifted_candles&quot;</span><span class="p">:</span><span class="w"> </span><span class="mi">2</span><span class="p">,</span>
<span class="w"> </span><span class="nt">&quot;indicator_periods_candles&quot;</span><span class="p">:</span><span class="w"> </span><span class="p">[</span><span class="mi">10</span><span class="p">,</span><span class="w"> </span><span class="mi">20</span><span class="p">]</span>
<span class="w"> </span><span class="p">},</span>
<span class="w"> </span><span class="c1">//...</span>
<span class="w"> </span><span class="p">}</span>
</code></pre></div>
<p>The <code>include_timeframes</code> in the config above are the timeframes (<code>tf</code>) of each call to <code>populate_any_indicators()</code> in the strategy. In the presented case, the user is asking for the <code>5m</code>, <code>15m</code>, and <code>4h</code> timeframes of the <code>rsi</code>, <code>mfi</code>, <code>roc</code>, and <code>bb_width</code> to be included in the feature set.</p>
<p>You can ask for each of the defined features to be included also for informative pairs using the <code>include_corr_pairlist</code>. This means that the feature set will include all the features from <code>populate_any_indicators</code> on all the <code>include_timeframes</code> for each of the correlated pairs defined in the config (<code>ETH/USD</code>, <code>LINK/USD</code>, and <code>BNB/USD</code> in the presented example).</p>
<p><code>include_shifted_candles</code> indicates the number of previous candles to include in the feature set. For example, <code>include_shifted_candles: 2</code> tells <code>FreqAI</code> to include the past 2 candles for each of the features in the feature set.</p>
<p>In total, the number of features the user of the presented example strat has created is: length of <code>include_timeframes</code> * no. features in <code>populate_any_indicators()</code> * length of <code>include_corr_pairlist</code> * no. <code>include_shifted_candles</code> * length of <code>indicator_periods_candles</code>
<span class="arithmatex">\(= 3 * 3 * 3 * 2 * 2 = 108\)</span>.</p>
<h3 id="returning-additional-info-from-training">Returning additional info from training<a class="headerlink" href="#returning-additional-info-from-training" title="Permanent link">&para;</a></h3>
<p>Important metrics can be returned to the strategy at the end of each model training by assigning them to <code>dk.data['extra_returns_per_train']['my_new_value'] = XYZ</code> inside the custom prediction model class. </p>
<p><code>FreqAI</code> takes the <code>my_new_value</code> assigned in this dictionary and expands it to fit the dataframe that is returned to the strategy. You can then use the returned metrics in your strategy through <code>dataframe['my_new_value']</code>. An example of how return values can be used in <code>FreqAI</code> are the <code>&amp;*_mean</code> and <code>&amp;*_std</code> values that are used to <a href="../freqai-configuration/#creating-a-dynamic-target-threshold">created a dynamic target threshold</a>.</p>
<p>Another example, where the user wants to use live metrics from the trade database, is shown below:</p>
<div class="highlight"><pre><span></span><code><span class="w"> </span><span class="nt">&quot;freqai&quot;</span><span class="p">:</span><span class="w"> </span><span class="p">{</span>
<span class="w"> </span><span class="nt">&quot;extra_returns_per_train&quot;</span><span class="p">:</span><span class="w"> </span><span class="p">{</span><span class="nt">&quot;total_profit&quot;</span><span class="p">:</span><span class="w"> </span><span class="mi">4</span><span class="p">}</span>
<span class="w"> </span><span class="p">}</span>
</code></pre></div>
<p>You need to set the standard dictionary in the config so that <code>FreqAI</code> can return proper dataframe shapes. These values will likely be overridden by the prediction model, but in the case where the model has yet to set them, or needs a default initial value, the preset values are what will be returned.</p>
<h2 id="feature-normalization">Feature normalization<a class="headerlink" href="#feature-normalization" title="Permanent link">&para;</a></h2>
<p><code>FreqAI</code> is strict when it comes to data normalization. The train features, <span class="arithmatex">\(X^{train}\)</span>, are always normalized to [-1, 1] using a shifted min-max normalization:</p>
<div class="arithmatex">\[X^{train}_{norm} = 2 * \frac{X^{train} - X^{train}.min()}{X^{train}.max() - X^{train}.min()} - 1\]</div>
<p>All other data (test data and unseen prediction data in dry/live/backtest) is always automatically normalized to the training feature space according to industry standards. <code>FreqAI</code> stores all the metadata required to ensure that test and prediction features will be properly normalized and that predictions are properly denormalized. For this reason, it is not recommended to eschew industry standards and modify <code>FreqAI</code> internals - however - advanced users can do so by inheriting <code>train()</code> in their custom <code>IFreqaiModel</code> and using their own normalization functions.</p>
<h2 id="data-dimensionality-reduction-with-principal-component-analysis">Data dimensionality reduction with Principal Component Analysis<a class="headerlink" href="#data-dimensionality-reduction-with-principal-component-analysis" title="Permanent link">&para;</a></h2>
<p>You can reduce the dimensionality of your features by activating the <code>principal_component_analysis</code> in the config:</p>
<div class="highlight"><pre><span></span><code><span class="w"> </span><span class="nt">&quot;freqai&quot;</span><span class="p">:</span><span class="w"> </span><span class="p">{</span>
<span class="w"> </span><span class="nt">&quot;feature_parameters&quot;</span><span class="w"> </span><span class="p">:</span><span class="w"> </span><span class="p">{</span>
<span class="w"> </span><span class="nt">&quot;principal_component_analysis&quot;</span><span class="p">:</span><span class="w"> </span><span class="kc">true</span>
<span class="w"> </span><span class="p">}</span>
<span class="w"> </span><span class="p">}</span>
</code></pre></div>
<p>This will perform PCA on the features and reduce their dimensionality so that the explained variance of the data set is &gt;= 0.999. Reducing data dimensionality makes training the model faster and hence allows for more up-to-date models. </p>
<h2 id="inlier-metric">Inlier metric<a class="headerlink" href="#inlier-metric" title="Permanent link">&para;</a></h2>
<p>The <code>inlier_metric</code> is a metric aimed at quantifying how similar a the features of a data point are to the most recent historic data points. </p>
<p>You define the lookback window by setting <code>inlier_metric_window</code> and <code>FreqAI</code> computes the distance between the present time point and each of the previous <code>inlier_metric_window</code> lookback points. A Weibull function is fit to each of the lookback distributions and its cumulative distribution function (CDF) is used to produce a quantile for each lookback point. The <code>inlier_metric</code> is then computed for each time point as the average of the corresponding lookback quantiles. The figure below explains the concept for an <code>inlier_metric_window</code> of 5.</p>
<p><img alt="inlier-metric" src="../assets/freqai_inlier-metric.jpg" /></p>
<p><code>FreqAI</code> adds the <code>inlier_metric</code> to the training features and hence gives the model access to a novel type of temporal information. </p>
<p>This function does <strong>not</strong> remove outliers from the data set.</p>
<h2 id="weighting-features-for-temporal-importance">Weighting features for temporal importance<a class="headerlink" href="#weighting-features-for-temporal-importance" title="Permanent link">&para;</a></h2>
<p><code>FreqAI</code> allows you to set a <code>weight_factor</code> to weight recent data more strongly than past data via an exponential function:</p>
<div class="arithmatex">\[ W_i = \exp(\frac{-i}{\alpha*n}) \]</div>
<p>where <span class="arithmatex">\(W_i\)</span> is the weight of data point <span class="arithmatex">\(i\)</span> in a total set of <span class="arithmatex">\(n\)</span> data points. Below is a figure showing the effect of different weight factors on the data points in a feature set.</p>
<p><img alt="weight-factor" src="../assets/freqai_weight-factor.jpg" /></p>
<h2 id="outlier-detection">Outlier detection<a class="headerlink" href="#outlier-detection" title="Permanent link">&para;</a></h2>
<p>Equity and crypto markets suffer from a high level of non-patterned noise in the form of outlier data points. <code>FreqAI</code> implements a variety of methods to identify such outliers and hence mitigate risk.</p>
<h3 id="identifying-outliers-with-the-dissimilarity-index-di">Identifying outliers with the Dissimilarity Index (DI)<a class="headerlink" href="#identifying-outliers-with-the-dissimilarity-index-di" title="Permanent link">&para;</a></h3>
<p>The Dissimilarity Index (DI) aims to quantify the uncertainty associated with each prediction made by the model. </p>
<p>You can tell <code>FreqAI</code> to remove outlier data points from the training/test data sets using the DI by including the following statement in the config:</p>
<div class="highlight"><pre><span></span><code><span class="w"> </span><span class="nt">&quot;freqai&quot;</span><span class="p">:</span><span class="w"> </span><span class="p">{</span>
<span class="w"> </span><span class="nt">&quot;feature_parameters&quot;</span><span class="w"> </span><span class="p">:</span><span class="w"> </span><span class="p">{</span>
<span class="w"> </span><span class="nt">&quot;DI_threshold&quot;</span><span class="p">:</span><span class="w"> </span><span class="mi">1</span>
<span class="w"> </span><span class="p">}</span>
<span class="w"> </span><span class="p">}</span>
</code></pre></div>
<p>The DI allows predictions which are outliers (not existent in the model feature space) to be thrown out due to low levels of certainty. To do so, <code>FreqAI</code> measures the distance between each training data point (feature vector), <span class="arithmatex">\(X_{a}\)</span>, and all other training data points:</p>
<div class="arithmatex">\[ d_{ab} = \sqrt{\sum_{j=1}^p(X_{a,j}-X_{b,j})^2} \]</div>
<p>where <span class="arithmatex">\(d_{ab}\)</span> is the distance between the normalized points <span class="arithmatex">\(a\)</span> and <span class="arithmatex">\(b\)</span>, and <span class="arithmatex">\(p\)</span> is the number of features, i.e., the length of the vector <span class="arithmatex">\(X\)</span>. The characteristic distance, <span class="arithmatex">\(\overline{d}\)</span>, for a set of training data points is simply the mean of the average distances:</p>
<div class="arithmatex">\[ \overline{d} = \sum_{a=1}^n(\sum_{b=1}^n(d_{ab}/n)/n) \]</div>
<p><span class="arithmatex">\(\overline{d}\)</span> quantifies the spread of the training data, which is compared to the distance between a new prediction feature vectors, <span class="arithmatex">\(X_k\)</span> and all the training data:</p>
<div class="arithmatex">\[ d_k = \arg \min d_{k,i} \]</div>
<p>This enables the estimation of the Dissimilarity Index as:</p>
<div class="arithmatex">\[ DI_k = d_k/\overline{d} \]</div>
<p>You can tweak the DI through the <code>DI_threshold</code> to increase or decrease the extrapolation of the trained model. A higher <code>DI_threshold</code> means that the DI is more lenient and allows predictions further away from the training data to be used whilst a lower <code>DI_threshold</code> has the opposite effect and hence discards more predictions.</p>
<p>Below is a figure that describes the DI for a 3D data set.</p>
<p><img alt="DI" src="../assets/freqai_DI.jpg" /></p>
<h3 id="identifying-outliers-using-a-support-vector-machine-svm">Identifying outliers using a Support Vector Machine (SVM)<a class="headerlink" href="#identifying-outliers-using-a-support-vector-machine-svm" title="Permanent link">&para;</a></h3>
<p>You can tell <code>FreqAI</code> to remove outlier data points from the training/test data sets using a Support Vector Machine (SVM) by including the following statement in the config:</p>
<div class="highlight"><pre><span></span><code><span class="w"> </span><span class="nt">&quot;freqai&quot;</span><span class="p">:</span><span class="w"> </span><span class="p">{</span>
<span class="w"> </span><span class="nt">&quot;feature_parameters&quot;</span><span class="w"> </span><span class="p">:</span><span class="w"> </span><span class="p">{</span>
<span class="w"> </span><span class="nt">&quot;use_SVM_to_remove_outliers&quot;</span><span class="p">:</span><span class="w"> </span><span class="kc">true</span>
<span class="w"> </span><span class="p">}</span>
<span class="w"> </span><span class="p">}</span>
</code></pre></div>
<p>The SVM will be trained on the training data and any data point that the SVM deems to be beyond the feature space will be removed.</p>
<p><code>FreqAI</code> uses <code>sklearn.linear_model.SGDOneClassSVM</code> (details are available on scikit-learn's webpage <a href="https://scikit-learn.org/stable/modules/generated/sklearn.linear_model.SGDOneClassSVM.html">here</a> (external website)) and you can elect to provide additional parameters for the SVM, such as <code>shuffle</code>, and <code>nu</code>.</p>
<p>The parameter <code>shuffle</code> is by default set to <code>False</code> to ensure consistent results. If it is set to <code>True</code>, running the SVM multiple times on the same data set might result in different outcomes due to <code>max_iter</code> being to low for the algorithm to reach the demanded <code>tol</code>. Increasing <code>max_iter</code> solves this issue but causes the procedure to take longer time.</p>
<p>The parameter <code>nu</code>, <em>very</em> broadly, is the amount of data points that should be considered outliers and should be between 0 and 1.</p>
<h3 id="identifying-outliers-with-dbscan">Identifying outliers with DBSCAN<a class="headerlink" href="#identifying-outliers-with-dbscan" title="Permanent link">&para;</a></h3>
<p>You can configure <code>FreqAI</code> to use DBSCAN to cluster and remove outliers from the training/test data set or incoming outliers from predictions, by activating <code>use_DBSCAN_to_remove_outliers</code> in the config:</p>
<div class="highlight"><pre><span></span><code><span class="w"> </span><span class="nt">&quot;freqai&quot;</span><span class="p">:</span><span class="w"> </span><span class="p">{</span>
<span class="w"> </span><span class="nt">&quot;feature_parameters&quot;</span><span class="w"> </span><span class="p">:</span><span class="w"> </span><span class="p">{</span>
<span class="w"> </span><span class="nt">&quot;use_DBSCAN_to_remove_outliers&quot;</span><span class="p">:</span><span class="w"> </span><span class="kc">true</span>
<span class="w"> </span><span class="p">}</span>
<span class="w"> </span><span class="p">}</span>
</code></pre></div>
<p>DBSCAN is an unsupervised machine learning algorithm that clusters data without needing to know how many clusters there should be.</p>
<p>Given a number of data points <span class="arithmatex">\(N\)</span>, and a distance <span class="arithmatex">\(\varepsilon\)</span>, DBSCAN clusters the data set by setting all data points that have <span class="arithmatex">\(N-1\)</span> other data points within a distance of <span class="arithmatex">\(\varepsilon\)</span> as <em>core points</em>. A data point that is within a distance of <span class="arithmatex">\(\varepsilon\)</span> from a <em>core point</em> but that does not have <span class="arithmatex">\(N-1\)</span> other data points within a distance of <span class="arithmatex">\(\varepsilon\)</span> from itself is considered an <em>edge point</em>. A cluster is then the collection of <em>core points</em> and <em>edge points</em>. Data points that have no other data points at a distance <span class="arithmatex">\(&lt;\varepsilon\)</span> are considered outliers. The figure below shows a cluster with <span class="arithmatex">\(N = 3\)</span>.</p>
<p><img alt="dbscan" src="../assets/freqai_dbscan.jpg" /></p>
<p><code>FreqAI</code> uses <code>sklearn.cluster.DBSCAN</code> (details are available on scikit-learn's webpage <a href="https://scikit-learn.org/stable/modules/generated/sklearn.cluster.DBSCAN.html">here</a> (external website)) with <code>min_samples</code> (<span class="arithmatex">\(N\)</span>) taken as &frac14; of the no. of time points in the feature set. <code>eps</code> (<span class="arithmatex">\(\varepsilon\)</span>) is computed automatically as the elbow point in the <em>k-distance graph</em> computed from the nearest neighbors in the pairwise distances of all data points in the feature set.</p>
</article>
</div>
</div>
</main>
<footer class="md-footer">
<nav class="md-footer__inner md-grid" aria-label="Footer" >
<a href="../freqai-parameter-table/" class="md-footer__link md-footer__link--prev" aria-label="Previous: Parameter table" rel="prev">
<div class="md-footer__button md-icon">
<svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 24 24"><path d="M20 11v2H8l5.5 5.5-1.42 1.42L4.16 12l7.92-7.92L13.5 5.5 8 11h12Z"/></svg>
</div>
<div class="md-footer__title">
<div class="md-ellipsis">
<span class="md-footer__direction">
Previous
</span>
Parameter table
</div>
</div>
</a>
<a href="../freqai-running/" class="md-footer__link md-footer__link--next" aria-label="Next: Running FreqAI" rel="next">
<div class="md-footer__title">
<div class="md-ellipsis">
<span class="md-footer__direction">
Next
</span>
Running FreqAI
</div>
</div>
<div class="md-footer__button md-icon">
<svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 24 24"><path d="M4 11v2h12l-5.5 5.5 1.42 1.42L19.84 12l-7.92-7.92L10.5 5.5 16 11H4Z"/></svg>
</div>
</a>
</nav>
<div class="md-footer-meta md-typeset">
<div class="md-footer-meta__inner md-grid">
<div class="md-copyright">
Made with
<a href="https://squidfunk.github.io/mkdocs-material/" target="_blank" rel="noopener">
Material for MkDocs
</a>
</div>
</div>
</div>
</footer>
<!-- Place this tag in your head or just before your close body tag. -->
<script async defer src="https://buttons.github.io/buttons.js"></script>
<script src="https://code.jquery.com/jquery-3.4.1.min.js"
integrity="sha256-CSXorXvZcTkaix6Yvo6HppcZGetbYMGWSFlBw8HfCJo=" crossorigin="anonymous"></script>
<!-- Load binance SDK -->
<script async defer src="https://public.bnbstatic.com/static/js/broker-sdk/broker-sdk@1.0.0.min.js"></script>
<script>
window.onload = function () {
var sidebar = document.getElementById('widget-wrapper')
var newDiv = document.createElement("div");
newDiv.id = "widget";
try {
sidebar.prepend(newDiv);
window.binanceBrokerPortalSdk.initBrokerSDK('#widget', {
apiHost: 'https://www.binance.com',
brokerId: 'R4BD3S82',
slideTime: 4e4,
});
} catch(err) {
console.log(err)
}
}
</script>
</div>
<div class="md-dialog" data-md-component="dialog">
<div class="md-dialog__inner md-typeset"></div>
</div>
<script id="__config" type="application/json">{"base": "..", "features": ["search.share"], "search": "../assets/javascripts/workers/search.5bf1dace.min.js", "translations": {"clipboard.copied": "Copied to clipboard", "clipboard.copy": "Copy to clipboard", "search.config.lang": "en", "search.config.pipeline": "trimmer, stopWordFilter", "search.config.separator": "[\\s\\-]+", "search.placeholder": "Search", "search.result.more.one": "1 more on this page", "search.result.more.other": "# more on this page", "search.result.none": "No matching documents", "search.result.one": "1 matching document", "search.result.other": "# matching documents", "search.result.placeholder": "Type to start searching", "search.result.term.missing": "Missing", "select.version.title": "Select version"}, "version": {"provider": "mike"}}</script>
<script src="../assets/javascripts/bundle.37e9125f.min.js"></script>
<script src="../javascripts/config.js"></script>
<script src="https://polyfill.io/v3/polyfill.min.js?features=es6"></script>
<script src="https://cdn.jsdelivr.net/npm/mathjax@3/es5/tex-mml-chtml.js"></script>
</body>
</html>
Reference in New Issue Copy Permalink