数学期望 Expectation

论坛 期权论坛     
匿名小用户   2019-10-20 17:54   55   0
<!-- flowchart 箭头图标 勿删 -->
                    <svg style="display: none;">
                        <path d="M5,0 0,2.5 5,5z" id="raphael-marker-block" stroke-linecap="round" style="-webkit-tap-highlight-color: rgba(0, 0, 0, 0);"></path>
                    </svg>
                                            <h1 id="数学期望-expectation">数学期望 Expectation</h1>



<h2 id="序言">序言</h2>

<p>机器学习中涉及到的很多概念都和 Expectation 相关联,例如:</p>

<ul>
<li>任何分布,我们都关心其 均值 mean、方差 variance、峰度 kurtosis、偏度 skewness;实际上都和数据期望相关;这些内容还和 中心距 以及 泰勒级数 相关联</li>
<li>机器学习中,针对模型的泛化能力 Generalization,常见对 泛化误差 decompose 为 Bias(偏差,欠拟合) 和 Variance(方差,过拟合)。这一过程就是利用 Expectation 推导的</li>
<li>最大期望算反 EM,明显地和 Expectation 有关联,其是利用了 Expectation 进行推导</li>
</ul>

<p>所以本文特此明确一下这个基础概念。</p>



<h2 id="期望的定义">期望的定义</h2>

<p>The average value of some function <span class="MathJax_Preview"></span><span aria-readonly="true" class="MathJax" id="MathJax-Element-1-Frame" role="textbox" style=""><nobr><span class="math" id="MathJax-Span-1" style="width: 2.017em; display: inline-block;"><span style="display: inline-block; position: relative; width: 1.6em; height: 0px; font-size: 125%;"><span style="position: absolute; clip: rect(1.722em, 1000em, 2.927em, -0.627em); top: -2.56em; left: 0em;"><span class="mrow" id="MathJax-Span-2"><span class="mi" id="MathJax-Span-3" style="font-family: STIXGeneral-Italic;">f<span style="display: inline-block; overflow: hidden; height: 1px; width: 0.146em;"></span></span><span class="mo" id="MathJax-Span-4" style="font-family: STIXGeneral-Regular;">(</span><span class="mi" id="MathJax-Span-5" style="font-family: STIXGeneral-Italic;">x<span style="display: inline-block; overflow: hidden; height: 1px; width: 0.003em;"></span></span><span class="mo" id="MathJax-Span-6" style="font-family: STIXGeneral-Regular;">)</span></span><span style="display: inline-block; width: 0px; height: 2.56em;"></span></span></span><span style="border-left-width: 0em; border-left-style: solid; display: inline-block; overflow: hidden; width: 0px; height: 1.24em; vertical-align: -0.325em;"></span></span></nobr></span> under a probability distribution <span class="MathJax_Preview"></span><span aria-readonly="true" class="MathJax" id="MathJax-Element-2-Frame" role="textbox" style=""><nobr><span class="math" id="MathJax-Span-7" style="width: 2.083em; display: inline-block;"><span style="display: inline-block; position: relative; width: 1.653em; height: 0px; font-size: 125%;"><span style="position: absolute; clip: rect(1.724em, 1000em, 2.925em, -0.555em); top: -2.56em; left: 0em;"><span class="mrow" id="MathJax-Span-8"><span class="mi" id="MathJax-Span-9" style="font-family: STIXGeneral-Italic;">p</span><span class="mo" id="MathJax-Span-10" style="font-family: STIXGeneral-Regular;">(</span><span class="mi" id="MathJax-Span-11" style="font-family: STIXGeneral-Italic;">x<span style="display: inline-block; overflow: hidden; height: 1px; width: 0.003em;"></span></span><span class="mo" id="MathJax-Span-12" style="font-family: STIXGeneral-Regular;">)</span></span><span style="display: inline-block; width: 0px; height: 2.56em;"></span></span></span><span style="border-left-width: 0em; border-left-style: solid; display: inline-block; overflow: hidden; width: 0px; height: 1.235em; vertical-align: -0.323em;"></span></span></nobr></span> is called the <strong>expectation</strong> of <span class="MathJax_Preview"></span><span aria-readonly="true" class="MathJax" id="MathJax-Element-3-Frame" role="textbox" style=""><nobr><span class="math" id="MathJax-Span-13" style="width: 2.017em; display: inline-block;"><span style="display: inline-block; position: relative; width: 1.6em; height: 0px; font-size: 125%;"><span style="position: absolute; clip: rect(1.722em, 1000em, 2.927em, -0.627em); top: -2.56em; left: 0em;"><span class="mrow" id="MathJax-Span-14"><span class="mi" id="MathJax-Span-15" style="font-family: STIXGeneral-Italic;">f<span style="display: inline-block; overflow: hidden; height: 1px; width: 0.146em;"></span></span><span class="mo" id="MathJax-Span-16" style="font-family: STIXGeneral-Regular;">(</span><span class="mi" id="MathJax-Span-17" style="font-family: STIXGeneral-Italic;">x<span style="display: inline-block; overflow: hidden; height: 1px; width: 0.003em;"></span></span><span class="mo" id="MathJax-Span-18" style="font-family: STIXGeneral-Regular;">)</span></span><span style="display: inline-block; width: 0px; height: 2.56em;"></span></span></span><span style="border-left-width: 0em; border-left-style: solid; display: inline-block; overflow: hidden; width: 0px; height: 1.24em; vertical-align: -0.325em;"></span></span></nobr></span>.  </p>

<p>如果 <span class="MathJax_Preview"></span><span aria-readonly="true" class="MathJax" id="MathJax-Element-4-Frame" role="textbox" style=""><nobr><span class="math" id="MathJax-Span-19" style="width: 2.083em; display: inline-block;"><span style="display: inline-block; position: relative; width: 1.653em; height: 0px; font-size: 125%;"><span style="position: absolute; clip: rect(1.724em, 1000em, 2.925em, -0.555em); top: -2.56em; left: 0em;"><
分享到 :
0 人收藏
您需要登录后才可以回帖 登录 | 立即注册

本版积分规则

积分:50
帖子:1079
精华:0
期权论坛 期权论坛
发布
内容

下载期权论坛手机APP