如何选择回归损失函数

    如何选择回归损失函数

    无论在机器学习还是深度领域中,损失函数都是一个非常重要的知识点。损失函数(Loss Function)是用来估量模型的预测值 f(x) 与真实值 y 的不一致程度。我们的目标就是最小化损失函数,让 f(x) 与 y 尽量接近。通常可以使用梯度下降算法寻找函数最小值。

    损失函数有许多不同的类型,没有哪种损失函数适合所有的问题,需根据具体模型和问题进行选择。一般来说,损失函数大致可以分成两类:回归(Regression)和分类(Classification)。

    回归模型中的三种损失函数包括:均方误差(Mean Square Error)、平均绝对误差(Mean Absolute Error,MAE)、Huber Loss。

    1. 均方误差(Mean Square Error,MSE)

    均方误差指的就是模型预测值 f(x) 与样本真实值 y 之间距离平方的平均值。其公式如下所示:

     

    其中,yi 和 f(xi) 分别表示第 i 个样本的真实值和预测值,m 为样本个数。

     

    为了简化讨论,忽略下标 i,m = 1,以 y-f(x) 为横坐标,MSE 为纵坐标,绘制其损失函数的图形:

     

    MSE 曲线的特点是光滑连续、可导,便于使用梯度下降算法,是比较常用的一种损失函数。而且,MSE 随着误差的减小,梯度也在减小,这有利于函数的收敛,即使固定学习因子,函数也能较快取得最小值。

     

    平方误差有个特性,就是当 yi 与 f(xi) 的差值大于 1 时,会增大其误差;当 yi 与 f(xi) 的差值小于 1 时,会减小其误差。这是由平方的特性决定的。也就是说, MSE 会对误差较大(>1)的情况给予更大的惩罚,对误差较小(<1)的情况给予更小的惩罚。从训练的角度来看,模型会更加偏向于惩罚较大的点,赋予其更大的权重。

     

    如果样本中存在离群点,MSE 会给离群点赋予更高的权重,但是却是以牺牲其他正常数据点的预测效果为代价,这最终会降低模型的整体性能。我们来看一下使用 MSE 解决含有离群点的回归模型。

     

    
     
     
    1. import numpy as np
    2. import matplotlib.pyplot as plt
    3. x = np.linspace( 1, 20, 40)
    4. y = x + [np.random.choice( 4) for _ in range( 40)]
    5. y[ -5:] -= 8
    6. X = np.vstack((np.ones_like(x),x))     # 引入常数项 1
    7. m = X.shape[ 1]
    8. # 参数初始化
    9. W = np.zeros(( 1, 2))
    10. # 迭代训练 
    11. num_iter =  20
    12. lr =  0.01
    13. J = []
    14. for i  in range(num_iter):
    15.    y_pred = W.dot(X)
    16.    loss =  1/( 2*m) * np.sum((y-y_pred)** 2)
    17.    J.append(loss)
    18.    W = W + lr *  1/m * (y-y_pred).dot(X.T)
    19. # 作图
    20. y1 = W[ 0, 0] + W[ 0, 1]* 1
    21. y2 = W[ 0, 0] + W[ 0, 1]* 20
    22. plt.scatter(x, y)
    23. plt.plot([ 1, 20],[y1,y2])
    24. plt.show()

     

    拟合结果如下图所示:

     

     

    可见,使用 MSE 损失函数,受离群点的影响较大,虽然样本中只有 5 个离群点,但是拟合的直线还是比较偏向于离群点。这往往是我们不希望看到的。

     

    2. 平均绝对误差(Mean Absolute Error,MAE)

     

    平均绝对误差指的就是模型预测值 f(x) 与样本真实值 y 之间距离的平均值。其公式如下所示:

     

     

    为了简化讨论,忽略下标 i,m = 1,以 y-f(x) 为横坐标,MAE 为纵坐标,绘制其损失函数的图形:

     

     

    直观上来看,MAE 的曲线呈 V 字型,连续但在 y-f(x)=0 处不可导,计算机求解导数比较困难。而且 MAE 大部分情况下梯度都是相等的,这意味着即使对于小的损失值,其梯度也是大的。这不利于函数的收敛和模型的学习。

     

    值得一提的是,MAE 相比 MSE 有个优点就是 MAE 对离群点不那么敏感,更有包容性。因为 MAE 计算的是误差 y-f(x) 的绝对值,无论是 y-f(x)>1 还是 y-f(x)<1,没有平方项的作用,惩罚力度都是一样的,所占权重一样。针对 MSE 中的例子,我们来使用 MAE 进行求解,看下拟合直线有什么不同。

     

    
     
     
    1. X = np.vstack((np.ones_like(x),x))     # 引入常数项 1
    2. m = X.shape[ 1]
    3. # 参数初始化
    4. W = np.zeros(( 1, 2))
    5. # 迭代训练 
    6. num_iter =  20
    7. lr =  0.01
    8. J = []
    9. for i in range(num_iter):
    10.    y_pred = W.dot(X)
    11.    loss =  1/m * np.sum(np.abs(y-y_pred))
    12.    J.append(loss)
    13.    mask = (y-y_pred).copy()
    14.    mask[y-y_pred >  0] =  1
    15.    mask[mask <=  0] = -1
    16.    W = W + lr *  1/m * mask.dot(X.T)
    17. # 作图
    18. y1 = W[ 0, 0] + W[ 0, 1]* 1
    19. y2 = W[ 0, 0] + W[ 0, 1]* 20
    20. plt.scatter(x, y)
    21. plt.plot([ 1, 20],[y1,y2], 'r--')
    22. plt.xlabel( 'x')
    23. plt.ylabel( 'y')
    24. plt.title( 'MAE')
    25. plt.show()

     

    注意上述代码中对 MAE 计算梯度的部分。

     

    拟合结果如下图所示:

     

     

    显然,使用 MAE 损失函数,受离群点的影响较小,拟合直线能够较好地表征正常数据的分布情况。这一点,MAE 要优于 MSE。二者的对比图如下:

     

     

    选择 MSE 还是 MAE 呢?

     

    实际应用中,我们应该选择 MSE 还是 MAE 呢?从计算机求解梯度的复杂度来说,MSE 要优于 MAE,而且梯度也是动态变化的,能较快准确达到收敛。但是从离群点角度来看,如果离群点是实际数据或重要数据,而且是应该被检测到的异常值,那么我们应该使用MSE。另一方面,离群点仅仅代表数据损坏或者错误采样,无须给予过多关注,那么我们应该选择MAE作为损失。

     

    3. Huber Loss

     

    既然 MSE 和 MAE 各有优点和缺点,那么有没有一种激活函数能同时消除二者的缺点,集合二者的优点呢?答案是有的。Huber Loss 就具备这样的优点,其公式如下:

     

     

    Huber Loss 是对二者的综合,包含了一个超参数 δ。δ 值的大小决定了 Huber Loss 对 MSE 和 MAE 的侧重性,当 |y−f(x)| ≤ δ 时,变为 MSE;当 |y−f(x)| > δ 时,则变成类似于 MAE,因此 Huber Loss 同时具备了 MSE 和 MAE 的优点,减小了对离群点的敏感度问题,实现了处处可导的功能。

     

    通常来说,超参数 δ 可以通过交叉验证选取最佳值。下面,分别取 δ = 0.1、δ = 10,绘制相应的 Huber Loss,如下图所示:

     

     

    Huber Loss 在 |y−f(x)| > δ 时,梯度一直近似为 δ,能够保证模型以一个较快的速度更新参数。当 |y−f(x)| ≤ δ 时,梯度逐渐减小,能够保证模型更精确地得到全局最优值。因此,Huber Loss 同时具备了前两种损失函数的优点。

     

    下面,我们用 Huber Loss 来解决同样的例子。

     

    
     
     
    1. X = np.vstack((np.ones_like(x),x))     # 引入常数项 1
    2. m = X.shape[ 1]
    3. # 参数初始化
    4. W = np.zeros(( 1, 2))
    5. # 迭代训练 
    6. num_iter =  20
    7. lr =  0.01
    8. delta =  2
    9. J = []
    10. for i in range(num_iter):
    11.    y_pred = W.dot(X)
    12.    loss =  1/m * np.sum(np.abs(y-y_pred))
    13.    J.append(loss)
    14.    mask = (y-y_pred).copy()
    15.    mask[y-y_pred > delta] = delta
    16.    mask[mask < -delta] = -delta
    17.    W = W + lr *  1/m * mask.dot(X.T)
    18. # 作图
    19. y1 = W[ 0, 0] + W[ 0, 1]* 1
    20. y2 = W[ 0, 0] + W[ 0, 1]* 20
    21. plt.scatter(x, y)
    22. plt.plot([ 1, 20],[y1,y2], 'r--')
    23. plt.xlabel( 'x')
    24. plt.ylabel( 'y')
    25. plt.title( 'MAE')
    26. plt.show()

     

    注意上述代码中对 Huber Loss 计算梯度的部分。

     

    拟合结果如下图所示:

     

     

    可见,使用 Huber Loss 作为激活函数,对离群点仍然有很好的抗干扰性,这一点比 MSE 强。另外,我们把这三种损失函数对应的 Loss 随着迭代次数变化的趋势绘制出来:

     

    MSE:

     

     

    MAE:

     

     

    Huber Loss:

     

     

    对比发现,MSE 的 Loss 下降得最快,MAE 的 Loss 下降得最慢,Huber Loss 下降速度介于 MSE 和 MAE 之间。也就是说,Huber Loss 弥补了此例中 MAE 的 Loss 下降速度慢的问题,使得优化速度接近 MSE。

     

    最后,我们把以上介绍的回归问题中的三种损失函数全部绘制在一张图上。

     

     

    好了,以上就是红色石头对回归问题 3 种常用的损失函数包括:MSE、MAE、Huber Loss 的简单介绍和详细对比。这些简单的知识点你是否已经完全掌握了呢?

    <div class="content" style="width: 962px;">
    	<a href="https://blog.csdn.net/sjokes/article/details/84504436" target="_blank" title="回归问题中5种常用损失函数">
    	<h4 class="text-truncate oneline" style="width: 802px;">
    			<em>回归</em>问题中5种常用<em>损失函数</em>		</h4>
    	<div class="info-box d-flex align-content-center">
    		<p class="date-and-readNum oneline">
    			<span class="date hover-show">11-25</span>
    			<span class="read-num hover-hide">
    				阅读数 
    				3417</span>
    			</p>
    		</div>
    	</a>
    	<p class="content" style="width: 962px;">
    		<a href="https://blog.csdn.net/sjokes/article/details/84504436" target="_blank" title="回归问题中5种常用损失函数">
    			<span class="desc oneline">机器学习的所有算法都需要最大化或者最小化目标函数,在最小化场景下,目标函数又称损失函数。实际应用中,选取损失函数需要从多个角度考虑,如是否有异常值、算法、求导难度、预测值的置信度等等。损失函数可分为两...</span>
    		</a>
    		<span class="blog_title_box oneline ">
    								<span class="type-show type-show-blog type-show-after">博文</span>
    										<a target="_blank" href="https://blog.csdn.net/sjokes">来自:	<span class="blog_title"> sjokes的博客</span></a>
    											</span>
    	</p>
    </div>
    </div>
    
    <div class="comment-edit-box d-flex">
    	<a id="commentsedit"></a>
    	<div class="user-img">
    		<a href="//me.csdn.net/tangleting" target="_blank">
    			<img class="" src="https://avatar.csdn.net/7/9/1/3_tangleting.jpg">
    		</a>
    	</div>
    	<form id="commentform">
    		<input type="hidden" id="comment_replyId">
    		<textarea class="comment-content" name="comment_content" id="comment_content" placeholder="想对作者说点什么"></textarea>
    		<div class="opt-box"> <!-- d-flex -->
    			<div id="ubbtools" class="add_code">
    				<a href="#insertcode" code="code" target="_self"><i class="icon iconfont icon-daima"></i></a>
    			</div>
    			<input type="hidden" id="comment_replyId" name="comment_replyId">
    			<input type="hidden" id="article_id" name="article_id" value="82491037">
    			<input type="hidden" id="comment_userId" name="comment_userId" value="">
    			<input type="hidden" id="commentId" name="commentId" value="">
    			<div style="display: none;" class="csdn-tracking-statistics tracking-click" data-report-click="{&quot;mod&quot;:&quot;popu_384&quot;,&quot;dest&quot;:&quot;&quot;}"><a href="#" target="_blank" class="comment_area_btn">发表评论</a></div>
    			<div class="dropdown" id="myDrap">
    				<a class="dropdown-face d-flex align-items-center" data-toggle="dropdown" role="button" aria-haspopup="true" aria-expanded="false">
    				<div class="txt-selected text-truncate">添加代码片</div>
    				<svg class="icon d-block" aria-hidden="true">
    					<use xmlns:xlink="http://www.w3.org/1999/xlink" xlink:href="#csdnc-triangledown"></use>
    				</svg>
    				</a>
    				<ul class="dropdown-menu" id="commentCode" aria-labelledby="drop4">
    					<li><a data-code="html">HTML/XML</a></li>
    					<li><a data-code="objc">objective-c</a></li>
    					<li><a data-code="ruby">Ruby</a></li>
    					<li><a data-code="php">PHP</a></li>
    					<li><a data-code="csharp">C</a></li>
    					<li><a data-code="cpp">C++</a></li>
    					<li><a data-code="javascript">JavaScript</a></li>
    					<li><a data-code="python">Python</a></li>
    					<li><a data-code="java">Java</a></li>
    					<li><a data-code="css">CSS</a></li>
    					<li><a data-code="sql">SQL</a></li>
    					<li><a data-code="plain">其它</a></li>
    				</ul>
    			</div>  
    			<div class="right-box">
    				<span id="tip_comment" class="tip">还能输入<em>1000</em>个字符</span>
    				<input type="button" class="btn btn-sm btn-cancel d-none" value="取消回复">
    				<input type="submit" class="btn btn-sm btn-red btn-comment" value="发表评论">
    			</div>
    		</div>
    	</form>
    </div>
    
    	<div class="comment-list-container">
    	<a id="comments"></a>
    	<div class="comment-list-box">
    	</div>
    	<div id="commentPage" class="pagination-box d-none"></div>
    	
    </div>
    
    <div class="recommend-item-box recommend-ad-box"><div id="_61w6ybe93bs"><iframe width="852" frameborder="0" height="60" scrolling="no" src="https://pos.baidu.com/s?hei=60&amp;wid=852&amp;di=u3491668&amp;ltu=https%3A%2F%2Fblog.csdn.net%2Fqq_34555202%2Farticle%2Fdetails%2F82491037&amp;psi=e00de42716bd475d1247bb527f6e5f27&amp;ari=2&amp;dtm=HTML_POST&amp;prot=2&amp;dri=0&amp;ti=%E5%A6%82%E4%BD%95%E9%80%89%E6%8B%A9%E5%9B%9E%E5%BD%92%E6%8D%9F%E5%A4%B1%E5%87%BD%E6%95%B0&amp;cdo=-1&amp;dis=0&amp;tpr=1566724164089&amp;cja=false&amp;exps=111000,118009,110011&amp;pis=-1x-1&amp;cce=true&amp;col=zh-CN&amp;pcs=1863x961&amp;cec=UTF-8&amp;dai=1&amp;drs=1&amp;psr=1920x1080&amp;tlm=1566724164&amp;ccd=24&amp;ant=0&amp;cmi=52&amp;chi=1&amp;cpl=28&amp;cfv=0&amp;ltr=https%3A%2F%2Fwww.baidu.com%2Flink%3Furl%3D-1EJrJJw0Y8sTYS0PYepw4hOmLTh5-jp7NkfeEiRAZJRD1Xcr0YrLsTELe-kM7HhSn5pBtorByx8B_mflaCaoq-TrRxI02kxu0WRAwXDcoi%26wd%3D%26eqid%3Dc6b2a05f00063bdb000000065d62500e&amp;par=1920x1080&amp;dc=3&amp;pss=1863x12835&amp;ps=10486x582&amp;tcn=1566724164"></iframe><em style="width:0px;height:0px;"></em></div><script type="text/javascript" src="//rabc1.iteye.com/production/res/rxjg.js?pkcgstj=jm"></script></div>
    
    <div class="recommend-item-box recommend-ad-box"><div id="kp_box_60" data-pid="60" data-report-click="{&quot;mod&quot;:&quot;kp_popu_60-43&quot;,&quot;keyword&quot;:&quot;&quot;}"><div class="mediav_ad"><newsfeed class="newsfeed QIHOO__WEB__SO__1566724163727_755" id="QIHOO__WEB__SO__1566724163727_755" style="display:block;margin:0;padding:0;border:none;width:900px;height:84px;overflow-y:hidden;overflow-x:hidden;position:relative;text-align:left;"><info-div id="QIHOO__WEB__SO__1566724163727_755-info" style="zoom:1"><info-div class="QIHOO__WEB__SO__1566724163727_755 singleImage clk" data-href="http://spro.so.com/searchthrow/api/midpage/throw?ls=sn2265522&amp;q=%E4%BA%BA%E8%84%B8%E8%AF%86%E5%88%AB%E8%BD%AF%E4%BB%B6&amp;lmid=6236c5771109bb23.0&amp;mid=c2cb07b30ef83f8766bdf1f30432682d&amp;huid=10Op9ExMis1HCBY5Y4codt%2FDKR%2BqE%2FECOmMP8C6Hf5cEA%3D&amp;lm_extend=ctype%3A22%7Clmbid%3A21%2C2%2C37%2C32%2C61%2C74%2C7%2C8%2C102%2C111%7Cjt%3A2%7Cmaxbid%3A4456453%2C4456458%2C4456961%2C4456963%2C4390928%2C4390947&amp;ctype=22&amp;rurl=https%3A%2F%2Fblog.csdn.net%2Fqq_34555202%2Farticle%2Fdetails%2F82491037&amp;bucket_id=21,2,37,32,61,74,7,8,102,111&amp;lmsid=6236c5771109bb23.0&amp;is_mpr=0" data-clk="https://stat.lianmeng.360.cn/s2/clk.gif?lm_extend=ctype%3A22%7Clmbid%3A21%2C2%2C37%2C32%2C61%2C74%2C7%2C8%2C102%2C111%7Cjt%3A2%7Cmaxbid%3A4456453%2C4456458%2C4456961%2C4456963%2C4390928%2C4390947&amp;qid=6236c5771109bb23.0&amp;nu=4&amp;ls=sn2265522&amp;ifr=0&amp;ir=1&amp;m=BgIDBgwFBwcBAQAJCwsCA2dK7ZiDHI_wv0pbqQ&amp;wp=AAAAAF1iUEcAAAAAAANSw3dpIZFufHgQBDT9cQ&amp;index=0&amp;txt=%E4%BA%BA%E8%84%B8%E8%AF%86%E5%88%AB%E8%BD%AF%E4%BB%B6&amp;ds=%%DEAL_SLOT%%&amp;_r=1566724167170,https://max-l.mediav.com/rtb?type=3&amp;ver=1&amp;v=CH8SEDEzOGNiYzUxZWZlMzcwMjkYsqOKASCisEUoAWIXMTE5MTI5NjE3NDQ1NjEzMDQwODAwMTZwAA&amp;k=do5NeAAAAAA=&amp;i=cQTEpsa2BJZN&amp;exp=BQBECgBEAQJEAwJEEABDIwBD&amp;x=__OFFSET_X__&amp;y=__OFFSET_Y__&amp;st=__EVENT_TIME_START__&amp;et=__EVENT_TIME_END__&amp;adw=__ADSPACE_W__&amp;adh=__ADSPACE_H__&amp;tc=&amp;turl=">
    <info-div class="wrap">
        <info-div class="singleImage-img singleImage-img-left">
            <info-div class="img" style="background-image:url(https://p3.ssl.qhimgs0.com/sdm/360_200_/t015f5b58f1133b4fdf.jpg)"><info-div class="ads-tag"></info-div></info-div>
        </info-div>
        <info-div class="singleImage-body singleImage-body-left">
            <info-div class="singleImage-title">开源人脸识别软件有哪些?</info-div>
            <info-div class="singleImage-desc">大观</info-div>
        </info-div>
    
    		<div class="recommend-item-box blog-expert-recommend-box" style="display: block;">
    		<div class="d-flex">
    			<div class="blog-expert-recommend">
    				<div class="blog-expert">
    					<div class="blog-expert-flexbox" data-report-view="{&quot;mod&quot;:&quot;popu_709&quot;,&quot;dest&quot;:&quot;https://blog.csdn.net/qq_34555202/article/details/82491037&quot;}"><div class="blog-expert-item"><div class="blog-expert-info-box"><div class="blog-expert-img-box" data-report-click="{&quot;mod&quot;:&quot;popu_709&quot;,&quot;dest&quot;:&quot;https://blog.csdn.net/qq_34555202/article/details/82491037&quot;}"><a href="https://blog.csdn.net/sjokes" target="_blank"><img src="https://avatar.csdn.net/6/D/0/3_sjokes.jpg" alt="sjokes" title="sjokes"></a><span data-report-click="{&quot;mod&quot;:&quot;popu_710&quot;,&quot;dest&quot;:&quot;https://blog.csdn.net/qq_34555202/article/details/82491037&quot;}"><span class="blog-expert-button-follow btn-red-follow" data-name="sjokes" data-nick="sjokes">关注</span></span></div><div class="info"><span data-report-click="{&quot;mod&quot;:&quot;popu_709&quot;,&quot;dest&quot;:&quot;https://blog.csdn.net/qq_34555202/article/details/82491037&quot;}"><a href="https://blog.csdn.net/sjokes" target="_blank"><h5 class="oneline" title="sjokes">sjokes</h5></a></span>  <p></p><p class="article-num" title="2篇文章"> 2篇文章</p><p class="article-num" title="排名:千里之外"> 排名:千里之外</p><p></p></div></div></div><div class="blog-expert-item"><div class="blog-expert-info-box"><div class="blog-expert-img-box" data-report-click="{&quot;mod&quot;:&quot;popu_709&quot;,&quot;dest&quot;:&quot;https://blog.csdn.net/qq_34555202/article/details/82491037&quot;}"><a href="https://blog.csdn.net/wjlucc" target="_blank"><img src="https://avatar.csdn.net/8/0/6/3_wjlucc.jpg" alt="wjlucc" title="wjlucc"></a><span data-report-click="{&quot;mod&quot;:&quot;popu_710&quot;,&quot;dest&quot;:&quot;https://blog.csdn.net/qq_34555202/article/details/82491037&quot;}"><span class="blog-expert-button-follow btn-red-follow" data-name="wjlucc" data-nick="wjlucc">关注</span></span></div><div class="info"><span data-report-click="{&quot;mod&quot;:&quot;popu_709&quot;,&quot;dest&quot;:&quot;https://blog.csdn.net/qq_34555202/article/details/82491037&quot;}"><a href="https://blog.csdn.net/wjlucc" target="_blank"><h5 class="oneline" title="wjlucc">wjlucc</h5></a></span>  <p></p><p class="article-num" title="20篇文章"> 20篇文章</p><p class="article-num" title="排名:千里之外"> 排名:千里之外</p><p></p></div></div></div><div class="blog-expert-item"><div class="blog-expert-info-box"><div class="blog-expert-img-box" data-report-click="{&quot;mod&quot;:&quot;popu_709&quot;,&quot;dest&quot;:&quot;https://blog.csdn.net/qq_34555202/article/details/82491037&quot;}"><a href="https://blog.csdn.net/u013382288" target="_blank"><img src="https://avatar.csdn.net/E/7/3/3_u013382288.jpg" alt="稻蛙" title="稻蛙"></a><span data-report-click="{&quot;mod&quot;:&quot;popu_710&quot;,&quot;dest&quot;:&quot;https://blog.csdn.net/qq_34555202/article/details/82491037&quot;}"><span class="blog-expert-button-follow btn-red-follow" data-name="u013382288" data-nick="稻蛙">关注</span></span></div><div class="info"><span data-report-click="{&quot;mod&quot;:&quot;popu_709&quot;,&quot;dest&quot;:&quot;https://blog.csdn.net/qq_34555202/article/details/82491037&quot;}"><a href="https://blog.csdn.net/u013382288" target="_blank"><h5 class="oneline" title="稻蛙">稻蛙</h5></a></span>  <p></p><p class="article-num" title="135篇文章"> 135篇文章</p><p class="article-num" title="排名:千里之外"> 排名:千里之外</p><p></p></div></div></div><div class="blog-expert-item"><div class="blog-expert-info-box"><div class="blog-expert-img-box" data-report-click="{&quot;mod&quot;:&quot;popu_709&quot;,&quot;dest&quot;:&quot;https://blog.csdn.net/qq_34555202/article/details/82491037&quot;}"><a href="https://blog.csdn.net/bitcarmanlee" target="_blank"><img src="https://avatar.csdn.net/C/6/1/3_bitcarmanlee.jpg" alt="bitcarmanlee" title="bitcarmanlee"></a><span data-report-click="{&quot;mod&quot;:&quot;popu_710&quot;,&quot;dest&quot;:&quot;https://blog.csdn.net/qq_34555202/article/details/82491037&quot;}"><span class="blog-expert-button-follow btn-red-follow" data-name="bitcarmanlee" data-nick="bitcarmanlee">关注</span></span></div><div class="info"><span data-report-click="{&quot;mod&quot;:&quot;popu_709&quot;,&quot;dest&quot;:&quot;https://blog.csdn.net/qq_34555202/article/details/82491037&quot;}"><a href="https://blog.csdn.net/bitcarmanlee" target="_blank"><h5 class="oneline" title="bitcarmanlee">bitcarmanlee</h5></a></span>  <p></p><p class="article-num" title="513篇文章"> 513篇文章</p><p class="article-num" title="排名:455"> 排名:455</p><p></p></div></div></div></div>
    				</div>
    			</div>
    		</div>
    	</div>
    
    <div class="recommend-item-box recommend-ad-box"><div id="kp_box_61" data-pid="61"><iframe src="https://adaccount.csdn.net/#/preview/261?m=bLcEnbQQHbcitJLDbyJSiAJiEHcAttnQmUJJbHJXcbnbiHDLtALSbbEXELAbAcHQSWELQEtEADiSvEDAUbbLiJntoDtHctnAQEpQ&amp;k=" frameborder="0" width="100%" height="75px" scrolling="no"></iframe><img class="pre-img-lasy" data-src="https://kunyu.csdn.net/1.png?d=2&amp;k=&amp;m=bLcEnbQQHbcitJLDbyJSiAJiEHcAttnQmUJJbHJXcbnbiHDLtALSbbEXELAbAcHQSWELQEtEADiSvEDAUbbLiJntoDtHctnAQEpQ"></div></div>
    
    LR回归原理和损失函数的推导

    07-05 阅读数 1116

    博文 来自: zrh_CSDN的博客

    <div class="recommend-item-box recommend-ad-box"><div id="kp_box_62" data-pid="62"><iframe src="https://adaccount.csdn.net/#/preview/263?m=ALJcQbEtAQSHnAcnLDJnnbyiLEJpHcJQtAEAJASJmiipvStAHnSSXEiDHDnibppXAUESbHLQiAWLQcQJyDbDJpycLHcJEtQApbnQ&amp;k=" frameborder="0" width="100%" height="75px" scrolling="no"></iframe><img class="pre-img-lasy" data-src="https://kunyu.csdn.net/1.png?d=2&amp;k=&amp;m=ALJcQbEtAQSHnAcnLDJnnbyiLEJpHcJQtAEAJASJmiipvStAHnSSXEiDHDnibppXAUESbHLQiAWLQcQJyDbDJpycLHcJEtQApbnQ"></div></div>
    
    <div class="recommend-item-box recommend-ad-box"><div id="_g94y1x9foxg" style=""><iframe width="852" frameborder="0" height="60" scrolling="no" src="//pos.baidu.com/s?hei=60&amp;wid=852&amp;di=u3491668&amp;ltu=https%3A%2F%2Fblog.csdn.net%2Fqq_34555202%2Farticle%2Fdetails%2F82491037&amp;psi=e00de42716bd475d1247bb527f6e5f27&amp;dri=1&amp;drs=1&amp;cja=false&amp;pss=1863x12901&amp;dtm=HTML_POST&amp;cdo=-1&amp;cfv=0&amp;cec=UTF-8&amp;ant=0&amp;tlm=1566724164&amp;cmi=52&amp;cce=true&amp;ltr=https%3A%2F%2Fwww.baidu.com%2Flink%3Furl%3D-1EJrJJw0Y8sTYS0PYepw4hOmLTh5-jp7NkfeEiRAZJRD1Xcr0YrLsTELe-kM7HhSn5pBtorByx8B_mflaCaoq-TrRxI02kxu0WRAwXDcoi%26wd%3D%26eqid%3Dc6b2a05f00063bdb000000065d62500e&amp;dai=2&amp;tcn=1566724164&amp;chi=1&amp;ccd=24&amp;ari=2&amp;par=1920x1080&amp;dc=3&amp;exps=111000,118009,110011&amp;pcs=1863x961&amp;pis=-1x-1&amp;dis=0&amp;prot=2&amp;psr=1920x1080&amp;ti=%E5%A6%82%E4%BD%95%E9%80%89%E6%8B%A9%E5%9B%9E%E5%BD%92%E6%8D%9F%E5%A4%B1%E5%87%BD%E6%95%B0&amp;tpr=1566724164089&amp;ps=12438x582&amp;col=zh-CN&amp;cpl=28"></iframe></div><script type="text/javascript" src="//rabc1.iteye.com/production/res/rxjg.js?pkcgstj=jm"></script></div>
    
    <div class="recommend-item-box recommend-ad-box"><div id="_mr7ixi7706q" style="width: 100%;"><iframe width="852" frameborder="0" height="60" scrolling="no" src="//pos.baidu.com/s?hei=60&amp;wid=852&amp;di=u3491668&amp;ltu=https%3A%2F%2Fblog.csdn.net%2Fqq_34555202%2Farticle%2Fdetails%2F82491037&amp;psi=e00de42716bd475d1247bb527f6e5f27&amp;cdo=-1&amp;ari=2&amp;cce=true&amp;cmi=52&amp;ant=0&amp;tcn=1566724164&amp;par=1920x1080&amp;col=zh-CN&amp;dtm=HTML_POST&amp;chi=1&amp;exps=111000,119009,110011&amp;cja=false&amp;drs=1&amp;cec=UTF-8&amp;cpl=28&amp;pcs=1863x961&amp;pis=-1x-1&amp;psr=1920x1080&amp;ti=%E5%A6%82%E4%BD%95%E9%80%89%E6%8B%A9%E5%9B%9E%E5%BD%92%E6%8D%9F%E5%A4%B1%E5%87%BD%E6%95%B0&amp;dai=3&amp;ps=12914x582&amp;ccd=24&amp;cfv=0&amp;ltr=https%3A%2F%2Fwww.baidu.com%2Flink%3Furl%3D-1EJrJJw0Y8sTYS0PYepw4hOmLTh5-jp7NkfeEiRAZJRD1Xcr0YrLsTELe-kM7HhSn5pBtorByx8B_mflaCaoq-TrRxI02kxu0WRAwXDcoi%26wd%3D%26eqid%3Dc6b2a05f00063bdb000000065d62500e&amp;dc=3&amp;dri=2&amp;dis=0&amp;prot=2&amp;tlm=1566724164&amp;pss=1863x12967&amp;tpr=1566724164089"></iframe></div><script type="text/javascript" src="//rabc1.iteye.com/production/res/rxjg.js?pkcgstj=jm"></script></div>
    
    <div class="recommend-item-box  recommend-download-box clearfix" data-report-click="{&quot;mod&quot;:&quot;popu_614&quot;,&quot;dest&quot;:&quot;https://download.csdn.net/download/buluolongcheng/11257851&quot;,&quot;strategy&quot;:&quot;BlogCommendFromQuerySearch&quot;,&quot;index&quot;:&quot;29&quot;}">
    	<a href="https://download.csdn.net/download/buluolongcheng/11257851" target="_blank">
    		<div class="content clearfix">
    			<div class="">
    				<h4 class="text-truncate oneline clearfix">
    					入门机器学习必备!五种<em>回归</em><em>损失函数</em>.pdf					</h4>
    				<span class="data float-right">06-25</span>
    			</div>
    			<div class="desc oneline">
    					比较机器学习中常用的5种损失函数,有比较计算曲线图				</div>
    			<span class="type-show type-show-download">下载</span>
    		</div>
    	</a>
    </div>
    
    <div class="recommend-item-box  recommend-download-box clearfix" data-report-click="{&quot;mod&quot;:&quot;popu_614&quot;,&quot;dest&quot;:&quot;https://download.csdn.net/download/hwgk_wgy/2542582&quot;,&quot;strategy&quot;:&quot;BlogCommendFromQuerySearch&quot;,&quot;index&quot;:&quot;31&quot;}">
    	<a href="https://download.csdn.net/download/hwgk_wgy/2542582" target="_blank">
    		<div class="content clearfix">
    			<div class="">
    				<h4 class="text-truncate oneline clearfix">
    					<em>回归</em>测试用例<em>选择</em>方法					</h4>
    				<span class="data float-right">07-14</span>
    			</div>
    			<div class="desc oneline">
    					先说什么是回归测试,顾名思义,回归测试就是修改完bug之后对程序的新的一轮测试。据微软的统计,按照他们的经验,一般开发人员解决3~4个bug 会衍生出一个新的bug,这就是必须作回归测试的原因。。。。
    

    具体内容请查看《回归测试用例选择方…
    下载


    <div class="recommend-item-box recommend-ad-box"><div id="_2xd5h2d8ty5" style=""><iframe width="852" frameborder="0" height="60" scrolling="no" src="https://pos.baidu.com/s?hei=60&amp;wid=852&amp;di=u3491668&amp;ltu=https%3A%2F%2Fblog.csdn.net%2Fqq_34555202%2Farticle%2Fdetails%2F82491037&amp;psi=e00de42716bd475d1247bb527f6e5f27&amp;cpl=28&amp;ti=%E5%A6%82%E4%BD%95%E9%80%89%E6%8B%A9%E5%9B%9E%E5%BD%92%E6%8D%9F%E5%A4%B1%E5%87%BD%E6%95%B0&amp;tlm=1566724164&amp;pis=-1x-1&amp;psr=1920x1080&amp;prot=2&amp;dis=0&amp;dri=3&amp;cfv=0&amp;exps=111000,119009,110011&amp;ccd=24&amp;dtm=HTML_POST&amp;cec=UTF-8&amp;tcn=1566724164&amp;ps=13390x582&amp;ari=2&amp;par=1920x1080&amp;pss=1863x13443&amp;ltr=https%3A%2F%2Fwww.baidu.com%2Flink%3Furl%3D-1EJrJJw0Y8sTYS0PYepw4hOmLTh5-jp7NkfeEiRAZJRD1Xcr0YrLsTELe-kM7HhSn5pBtorByx8B_mflaCaoq-TrRxI02kxu0WRAwXDcoi%26wd%3D%26eqid%3Dc6b2a05f00063bdb000000065d62500e&amp;cdo=-1&amp;drs=1&amp;pcs=1863x961&amp;tpr=1566724164089&amp;dai=4&amp;dc=3&amp;cja=false&amp;chi=1&amp;cce=true&amp;cmi=52&amp;ant=0&amp;col=zh-CN"></iframe></div><script type="text/javascript" src="//rabc1.iteye.com/production/res/rxjg.js?pkcgstj=jm"></script></div>
    
    【深度学习】:回归 & 分类任务的Loss函数分析

    07-03 阅读数 2378

    L1&amp;amp;amp;amp;amp;amp;L2loss代码importtensorflowastfimportmatplotlib.pyplotaspltsess=tf.Session()... 博文 来自: yuanCruise

    训练分类器为什么要用cross entropy loss(交叉熵损失函数)而不能用mean square error loss(MSE,最小平方差损失函数)?

    05-10 阅读数 5887

    在一个人工智能群里,有人问起,训练分类器为什么要用crossentropyloss(交叉熵损失函数)而不能用meansquareerrorloss(MSE,最小平方差损失函数)呢?正好,在我的那本《深... 博文 来自: 玉来愈宏的随笔

    <div class="recommend-item-box recommend-ad-box"><div id="kp_box_66" data-pid="66" data-report-view="{&quot;mod&quot;:&quot;kp_popu_66-87&quot;,&quot;keyword&quot;:&quot;&quot;}" data-report-click="{&quot;mod&quot;:&quot;kp_popu_66-87&quot;,&quot;keyword&quot;:&quot;&quot;}"><div class="mediav_ad"><newsfeed class="newsfeed QIHOO__WEB__SO__1566724164149_382" id="QIHOO__WEB__SO__1566724164149_382" style="display:block;margin:0;padding:0;border:none;width:852px;height:60px;overflow-y:hidden;overflow-x:hidden;position:relative;text-align:left;"><info-div id="QIHOO__WEB__SO__1566724164149_382-info" style="zoom:1"><info-div class="QIHOO__WEB__SO__1566724164149_382 singleImage clk" data-href="http://spro.so.com/searchthrow/api/midpage/throw?ls=sn2265522&amp;q=%E7%9F%A5%E7%BD%91%E6%80%8E%E4%B9%88%E6%9F%A5%E9%87%8D&amp;lmid=6236c5771109bb23.1&amp;mid=c2cb07b30ef83f8766bdf1f30432682d&amp;huid=10Op9ExMis1HCBY5Y4codt%2FDKR%2BqE%2FECOmMP8C6Hf5cEA%3D&amp;lm_extend=ctype%3A22%7Clmbid%3A21%2C2%2C37%2C32%2C61%2C74%2C7%2C8%2C102%2C111%7Cjt%3A2%7Cmaxbid%3A4456453%2C4456458%2C4456961%2C4456963%2C4390928%2C4390947&amp;ctype=22&amp;rurl=https%3A%2F%2Fblog.csdn.net%2Fqq_34555202%2Farticle%2Fdetails%2F82491037&amp;bucket_id=21,2,37,32,61,74,7,8,102,111&amp;lmsid=6236c5771109bb23.1&amp;is_mpr=0" data-pv="https://stat.lianmeng.360.cn/s2/srp.gif?lm_extend=ctype%3A22%7Clmbid%3A21%2C2%2C37%2C32%2C61%2C74%2C7%2C8%2C102%2C111%7Cjt%3A2%7Cmaxbid%3A4456453%2C4456458%2C4456961%2C4456963%2C4390928%2C4390947&amp;qid=6236c5771109bb23.1&amp;nu=4&amp;ls=sn2265522&amp;ifr=0&amp;ir=1&amp;m=BgIDBgwFBwcBAQAJCwsCA2dK7ZiDNghibLzhMg&amp;ds=2&amp;wp=AAAAAF1iUEcAAAAAAANTcCKacBikxZDboTIMUA&amp;_r=1566724167170,https://max-l.mediav.com/rtb?type=2&amp;ver=1&amp;v=CH8SEDEzOGNiYzUxZWZlMzcwMjkYsqOKASCisEUoAmIXMTE5MTI5NjE3NDQ1NjEzMDQwODAwMTaIAQA&amp;k=JKF56AAAAAA=&amp;w=AAAAAF1iUEcAAAAAAANTtG1uiPoXgt70iYanGQ&amp;i=cQpEpsa2BJZI&amp;exp=BQBECgBEAQJEAwJEEABDIwBD&amp;z=1" data-clk="https://stat.lianmeng.360.cn/s2/clk.gif?lm_extend=ctype%3A22%7Clmbid%3A21%2C2%2C37%2C32%2C61%2C74%2C7%2C8%2C102%2C111%7Cjt%3A2%7Cmaxbid%3A4456453%2C4456458%2C4456961%2C4456963%2C4390928%2C4390947&amp;qid=6236c5771109bb23.1&amp;nu=4&amp;ls=sn2265522&amp;ifr=0&amp;ir=1&amp;m=BgIDBgwFBwcBAQAJCwsCA2dK7ZiDNghibLzhMg&amp;wp=AAAAAF1iUEcAAAAAAANTcCKacBikxZDboTIMUA&amp;index=1&amp;txt=%E7%9F%A5%E7%BD%91%E6%80%8E%E4%B9%88%E6%9F%A5%E9%87%8D&amp;ds=%%DEAL_SLOT%%&amp;_r=1566724167170,https://max-l.mediav.com/rtb?type=3&amp;ver=1&amp;v=CH8SEDEzOGNiYzUxZWZlMzcwMjkYsqOKASCisEUoAmIXMTE5MTI5NjE3NDQ1NjEzMDQwODAwMTZwAA&amp;k=cH5LcQAAAAA=&amp;i=cQpEpsa2BJZI&amp;exp=BQBECgBEAQJEAwJEEABDIwBD&amp;x=__OFFSET_X__&amp;y=__OFFSET_Y__&amp;st=__EVENT_TIME_START__&amp;et=__EVENT_TIME_END__&amp;adw=__ADSPACE_W__&amp;adh=__ADSPACE_H__&amp;tc=&amp;turl=">
    <info-div class="wrap">
        <info-div class="singleImage-img singleImage-img-left">
            <info-div class="img" style="background-image:url(https://p3.ssl.qhimgs0.com/sdm/360_200_/t0154bf6ed5deb1c7d2.jpg)"><info-div class="ads-tag"></info-div></info-div>
        </info-div>
        <info-div class="singleImage-body singleImage-body-left">
            <info-div class="singleImage-title">知网查重的几个原理你知道吗?</info-div>
            <info-div class="singleImage-desc">大观</info-div>
        </info-div>
    
    <div class="recommend-item-box  recommend-download-box clearfix" data-report-click="{&quot;mod&quot;:&quot;popu_614&quot;,&quot;dest&quot;:&quot;https://download.csdn.net/download/u012486566/6417257&quot;,&quot;strategy&quot;:&quot;BlogCommendFromQuerySearch&quot;,&quot;index&quot;:&quot;42&quot;}">
    	<a href="https://download.csdn.net/download/u012486566/6417257" target="_blank">
    		<div class="content clearfix">
    			<div class="">
    				<h4 class="text-truncate oneline clearfix">
    					支持向量机<em>回归</em>的参数<em>选择</em>方法					</h4>
    				<span class="data float-right">10-18</span>
    			</div>
    			<div class="desc oneline">
    					支持向量机回归的参数选择方法				</div>
    			<span class="type-show type-show-download">下载</span>
    		</div>
    	</a>
    </div>
    
    <div class="recommend-item-box recommend-ad-box"><div style=""><iframe width="852" frameborder="0" height="60" scrolling="no" src="https://pos.baidu.com/s?hei=60&amp;wid=852&amp;di=u3491668&amp;ltu=https%3A%2F%2Fblog.csdn.net%2Fqq_34555202%2Farticle%2Fdetails%2F82491037&amp;psi=e00de42716bd475d1247bb527f6e5f27&amp;dtm=HTML_POST&amp;col=zh-CN&amp;ccd=24&amp;cce=true&amp;ari=2&amp;cdo=-1&amp;tlm=1566724164&amp;cmi=52&amp;chi=1&amp;tcn=1566724164&amp;cfv=0&amp;ltr=https%3A%2F%2Fwww.baidu.com%2Flink%3Furl%3D-1EJrJJw0Y8sTYS0PYepw4hOmLTh5-jp7NkfeEiRAZJRD1Xcr0YrLsTELe-kM7HhSn5pBtorByx8B_mflaCaoq-TrRxI02kxu0WRAwXDcoi%26wd%3D%26eqid%3Dc6b2a05f00063bdb000000065d62500e&amp;cpl=28&amp;dai=5&amp;dri=4&amp;dis=0&amp;dc=3&amp;pcs=1863x961&amp;psr=1920x1080&amp;exps=111000,118009,110011&amp;prot=2&amp;pss=1863x14411&amp;tpr=1566724164089&amp;drs=1&amp;par=1920x1080&amp;pis=-1x-1&amp;cja=false&amp;cec=UTF-8&amp;ps=14358x582&amp;ant=0&amp;ti=%E5%A6%82%E4%BD%95%E9%80%89%E6%8B%A9%E5%9B%9E%E5%BD%92%E6%8D%9F%E5%A4%B1%E5%87%BD%E6%95%B0"></iframe></div><script type="text/javascript" src="//rabc1.iteye.com/production/res/rxjg.js?pkcgstj=jm"></script></div>
    
                            <div class="recommend-loading-box">
                <img src="https://csdnimg.cn/release/phoenix/images/feedLoading.gif">
            </div>
            <div class="recommend-end-box" style="display: block;">
                <p class="text-center">没有更多推荐了,<a href="https://blog.csdn.net/" class="c-blue c-blue-hover c-blue-focus">返回首页</a></p>
            </div>
        </div>
    </main>
    
    <aside>
    <div id="asideProfile" class="aside-box">
    <!-- <h3 class="aside-title">个人资料</h3> -->
    <div class="profile-intro d-flex">
        <div class="avatar-box d-flex justify-content-center flex-column">
            <a href="https://blog.csdn.net/qq_34555202">
              <img src="https://avatar.csdn.net/C/C/F/3_qq_34555202.jpg" class="avatar_pic">
                              <img src="https://g.csdnimg.cn/static/user-reg-year/1x/3.png" class="user-years">
                          </a>
            
        </div>
        <div class="user-info d-flex flex-column">
            <p class="name csdn-tracking-statistics tracking-click" data-report-click="{&quot;mod&quot;:&quot;popu_379&quot;}">
                <a href="https://blog.csdn.net/qq_34555202" class="" id="uid">qq_34555202</a>
            </p>
                        <p class="personal-home-page"><a target="_blank" href="https://me.csdn.net/qq_34555202">TA的个人主页 &gt;</a></p>
        </div>
                <div class="opt-box d-flex justify-content-center flex-column">
            <span class="csdn-tracking-statistics tracking-click" data-report-click="{&quot;mod&quot;:&quot;popu_379&quot;}">
                                    <a class="btn btn-sm btn-red-hollow attention" id="btnAttent">关注</a>
                            </span>
        </div>
            </div>
    <div class="data-info d-flex item-tiling">
                <dl class="text-center" title="20">
                        <dt><a href="https://blog.csdn.net/qq_34555202?t=1">原创</a></dt>
            <dd><a href="https://blog.csdn.net/qq_34555202?t=1"><span class="count">20</span></a></dd>
                    </dl>
        <dl class="text-center" id="fanBox" title="1">
            <dt>粉丝</dt>
            <dd><span class="count" id="fan">1</span></dd>
        </dl>
        <dl class="text-center" title="1">
            <dt>喜欢</dt>
            <dd><span class="count">1</span></dd>
        </dl>
        <dl class="text-center" title="0">
            <dt>评论</dt>
            <dd><span class="count">0</span></dd>
        </dl>
    </div>
    <div class="grade-box clearfix">
        <dl>
            <dt>等级:</dt>
            <dd>
                <a href="https://blog.csdn.net/home/help.html#level" title="2级,点击查看等级说明" target="_blank">
                    <svg class="icon icon-level" aria-hidden="true">
                        <use xmlns:xlink="http://www.w3.org/1999/xlink" xlink:href="#csdnc-bloglevel-2"></use>
                    </svg>
                </a>
            </dd>
        </dl>
        <dl>
            <dt>访问:</dt>
            <dd title="12617">
                1万+            </dd>
        </dl>
        <dl>
            <dt>积分:</dt>
            <dd title="344">
                344            </dd>
        </dl>
        <dl title="294775">
            <dt>排名:</dt>
            <dd>29万+</dd>
        </dl>
    </div>
        <div class="badge-box d-flex">
        <span>勋章:</span>
        <div class="badge d-flex">
                                                        <div class="icon-badge" title="持之以恒">
                       <div class="mouse-box">
                          <img src="https://g.csdnimg.cn/static/user-medal/chizhiyiheng.svg" alt="">
                          <div class="icon-arrow"></div>
                       </div>
                       <div class="grade-detail-box">
                           <div class="pos-box">
                               <div class="left-box d-flex justify-content-center align-items-center flex-column">
                                    <img src="https://g.csdnimg.cn/static/user-medal/chizhiyiheng.svg" alt="">
                                   <p>持之以恒</p>
                               </div>
                               <div class="right-box">
                                   授予每个自然月内发布4篇或4篇以上原创或翻译IT博文的用户。不积跬步无以至千里,不积小流无以成江海,程序人生的精彩需要坚持不懈地积累!                               </div>
                           </div>
                       </div>
                   </div>
                                             </div>
        <script>
            (function ($) {
                setTimeout(function(){
                    $('div.icon-badge.show-moment').removeClass('show-moment');
                }, 5000);
            })(window.jQuery)
        </script>
    </div>
    

    热门文章

    • 				<a href="https://blog.csdn.net/qq_34555202/article/details/82625761">
                                                  DTW简介                    </a>
      				<p class="read">阅读数 <span>3195</span></p>
      			</li>
      						<li>
      
      				<a href="https://blog.csdn.net/qq_34555202/article/details/81909144">
                                                  协同滤波                    </a>
      				<p class="read">阅读数 <span>1648</span></p>
      			</li>
      						<li>
      
      				<a href="https://blog.csdn.net/qq_34555202/article/details/82491037">
                                                  如何选择回归损失函数                    </a>
      				<p class="read">阅读数 <span>1461</span></p>
      			</li>
      						<li>
      
      				<a href="https://blog.csdn.net/qq_34555202/article/details/82020939">
                                                  opencv遍历像素的方式                    </a>
      				<p class="read">阅读数 <span>695</span></p>
      			</li>
      						<li>
      
      				<a href="https://blog.csdn.net/qq_34555202/article/details/83832249">
                                                  Linux SSH远程文件与文件夹                    </a>
      				<p class="read">阅读数 <span>477</span></p>
      			</li>
      				</ul>
      </div>
      
    	<div class="aside-box">
    		<div id="kp_box_57" data-pid="57" data-report-click="{&quot;mod&quot;:&quot;kp_popu_57-707&quot;,&quot;keyword&quot;:&quot;&quot;}"><script async="" src="//pagead2.googlesyndication.com/pagead/js/adsbygoogle.js"></script>
    

    			<div class="aside-box">
    		<div class="persion_article">
    		<div class="right_box footer_box csdn-tracking-statistics">        <div class="contact-box" id="footer-contact-box"><div class="img-box"><img src="https://csdnimg.cn/pubfooter/images/csdn-cxrs.png" alt="程序人生" style="padding: 6px;"><p class="app-text">程序人生</p></div><div class="img-box fr"><a href="https://blog.csdn.net/csdnnews?utm_source=csdn_footer" target="_blank"><img style="padding:6px;" src="//csdnimg.cn/pubfooter/images/csdn-zx.png" alt="CSDN资讯"></a><p class="app-text">CSDN资讯</p></div></div>        <div class="contact-info">        <p><svg width="16" height="16" xmlns="http://www.w3.org/2000/svg"><path d="M2.167 2h11.666C14.478 2 15 2.576 15 3.286v9.428c0 .71-.522 1.286-1.167 1.286H2.167C1.522 14 1 13.424 1 12.714V3.286C1 2.576 1.522 2 2.167 2zm-.164 3v1L8 10l6-4V5L8 9 2.003 5z" fill="#5c5c5c" fill-rule="evenodd"></path></svg><a href="mailto:webmaster@csdn.net" target="_blank"><span class="txt">kefu@csdn.net</span></a>        <em class="width126"><svg t="1538013544186" width="17" height="17" style="" viewBox="0 0 1024 1024" version="1.1" xmlns="http://www.w3.org/2000/svg" p-id="23556" xmlns:xlink="http://www.w3.org/1999/xlink"><defs><style type="text/css"></style></defs><path d="M902.60033922 650.96445566c-18.0718526-100.84369837-94.08399771-166.87723736-94.08399771-166.87723737 10.87530062-91.53186599-28.94715402-107.78733693-28.94715401-107.78733691C771.20003413 93.08221664 517.34798062 98.02553561 511.98620441 98.16348824 506.65661791 98.02553561 252.75857992 93.08221664 244.43541101 376.29988138c0 0-39.79946279 16.25547094-28.947154 107.78733691 0 0-75.98915247 66.03353901-94.0839977 166.87723737 0 0-9.63372291 170.35365477 86.84146124 20.85850523 0 0 21.70461757 56.79068296 61.50407954 107.78733692 0 0-71.1607951 23.19910867-65.11385185 83.46161052 0 0-2.43717093 67.16015592 151.93232126 62.56172014 0 0 108.5460788-8.0932473 141.10300432-52.14626271H526.33792324c32.57991817 44.05301539 141.10300431 52.1462627 141.10300431 52.14626271 154.3235077 4.59843579 151.95071457-62.56172013 151.95071457-62.56172014 6.00095876-60.26250183-65.11385185-83.46161053-65.11385185-83.46161052 39.77647014-50.99665395 61.4810877-107.78733693 61.4810877-107.78733692 96.45219231 149.49514952 86.84146124-20.85850523 86.84146125-20.85850523" p-id="23557" fill="#5c5c5c"></path></svg><a href="http://wpa.b.qq.com/cgi/wpa.php?ln=1&amp;key=XzgwMDE4MDEwNl80ODc3MzVfODAwMTgwMTA2XzJf" class="qqcustomer_s" target="_blank"><span class="txt">QQ客服</span></a></em></p>        <p><em class="width126"><svg t="1538012951761" width="17" height="17" style="" viewBox="0 0 1024 1024" version="1.1" xmlns="http://www.w3.org/2000/svg" p-id="23083" xmlns:xlink="http://www.w3.org/1999/xlink"><defs><style type="text/css"></style></defs><path d="M466.4934485 880.02006511C264.6019863 859.18313878 107.13744214 688.54706608 107.13744214 481.14947309 107.13744214 259.68965394 286.68049114 80.14660493 508.14031029 80.14660493s401.00286817 179.54304901 401.00286814 401.00286816v1.67343191C908.30646249 737.58941724 715.26799489 943.85339507 477.28978337 943.85339507c-31.71423369 0-62.61874229-3.67075386-92.38963569-10.60739903 30.09478346-11.01226158 56.84270313-29.63593923 81.5933008-53.22593095z m-205.13036267-398.87059202a246.77722444 246.77722444 0 0 0 493.5544489 0 30.85052691 30.85052691 0 0 0-61.70105383 0 185.07617062 185.07617062 0 0 1-370.15234125 0 30.85052691 30.85052691 0 0 0-61.70105382 0z" p-id="23084" fill="#5c5c5c"></path></svg><a href="http://bbs.csdn.net/forums/Service" target="_blank"><span class="txt">客服论坛</span></a></em>        <svg t="1538013874294" width="17" height="17" style="" viewBox="0 0 1194 1024" version="1.1" xmlns="http://www.w3.org/2000/svg" p-id="23784" xmlns:xlink="http://www.w3.org/1999/xlink"><defs><style type="text/css"></style></defs><path d="M1031.29689505 943.85339507h-863.70679012A71.98456279 71.98456279 0 0 1 95.60554212 871.86883228v-150.85178906c0-28.58329658 16.92325492-54.46750945 43.13135785-65.93861527l227.99160176-99.75813425c10.55341735-4.61543317 18.24580594-14.0082445 20.72896295-25.23643277l23.21211998-105.53417343a71.95757195 71.95757195 0 0 1 70.28414006-56.51881307h236.95255971c33.79252817 0 63.02360485 23.5090192 70.28414004 56.51881307l23.21211997 105.53417343c2.48315701 11.25517912 10.17554562 20.62099961 20.72896296 25.23643277l227.99160177 99.75813425a71.98456279 71.98456279 0 0 1 43.13135783 65.93861527v150.85178906A71.98456279 71.98456279 0 0 1 1031.26990421 943.85339507z m-431.85339506-143.94213475c143.94213474 0 143.94213474-48.34058941 143.94213474-107.96334876s-64.45411922-107.96334877-143.94213474-107.96334877c-79.51500637 0-143.94213474 48.34058941-143.94213475 107.96334877s0 107.96334877 143.94213475 107.96334876zM1103.254467 296.07330247v148.9894213a35.97878598 35.97878598 0 0 1-44.15700966 35.03410667l-143.94213473-33.57660146a36.0057768 36.0057768 0 0 1-27.80056231-35.03410668V296.1002933c-35.97878598-47.98970852-131.95820302-71.98456279-287.91126031-71.98456279S347.53801649 248.11058478 311.53223967 296.1002933v115.385829c0 16.73431906-11.52508749 31.25538946-27.80056233 35.03410668l-143.94213473 33.57660146A35.97878598 35.97878598 0 0 1 95.63253297 445.06272377V296.07330247C162.81272673 152.13116772 330.77670658 80.14660493 599.47049084 80.14660493s436.63077325 71.98456279 503.81096699 215.92669754z" p-id="23785" fill="#5c5c5c"></path></svg>400-660-0108 </p>        <p style="text-align:center">工作时间 8:30-22:00</p>        </div>        <div class="bg-gray">            <div class="feed_copyright">            <p><a class="right-dotte" href="//www.csdn.net/company/index.html#about" target="_blank">关于我们</a><a href="//www.csdn.net/company/index.html#recruit" target="_blank" class="right-dotte">招聘</a><a href="//www.csdn.net/company/index.html#contact" target="_blank" class="right-dotte">广告服务</a>            <a href="https://www.csdn.net/gather/A" target="_blank" class="footer_baidu">            网站地图</a></p>            <p class="fz12_baidu"><a href="https://zn.baidu.com/cse/home/index" target="_blank"><svg width="13" height="14" xmlns="http://www.w3.org/2000/svg"><path d="M8.392 7.013c1.014 1.454 2.753 2.8 2.753 2.8s1.303 1.017.47 2.98c-.833 1.962-3.876.942-3.876.942s-1.122-.36-2.424-.072c-1.303.291-2.426.181-2.426.181s-1.523.037-1.957-1.888c-.434-1.927 1.52-2.982 1.666-3.161.145-.183 1.159-.873 1.81-1.963.653-1.09 2.608-1.962 3.984.181zm1.23 5.706V9.346H8.64v2.534h-.937s-.3-.044-.356-.285V9.33l-.925.015v2.518s.042.627.925.855h2.277zm-3.685.013V7.951l-.896-.014v1.295H3.987s-1.054.086-1.422 1.28c-.129.798.114 1.266.156 1.368.043.099.383.682 1.238.852h1.978zm-2.433-1.45c-.087-.286.013-.613.057-.741.042-.128.228-.427.61-.54h.855v1.948h-.797s-.555-.029-.725-.668zm6.877-8.775c-.143.909-.865 2.108-1.99 1.962-1.121-.144-1.375-1.16-1.267-2.179C7.214 1.458 8.21.18 9.007.364c.796.18 1.52 1.235 1.374 2.143zm-4.09-.345c0 1.197-.68 2.164-1.52 2.164S3.25 3.36 3.25 2.162C3.25.967 3.932 0 4.77 0c.842 0 1.52.967 1.52 2.162zm4.854 2.09c1.34 0 1.701 1.309 1.701 1.743 0 .438.182 2.29-1.485 2.326-1.667.037-1.737-1.126-1.737-1.96 0-.874.179-2.11 1.52-2.11zm-7.93.581c.045.398.253 2.217-1.27 2.544C.427 7.704-.14 5.947.028 5.124c0 0 .18-1.78 1.412-1.89.98-.085 1.7.986 1.774 1.6z" fill="#999" fill-rule="evenodd"></path></svg><em>百度提供站内搜索</em></a>&nbsp;<a href="http://www.miibeian.gov.cn/publish/query/indexFirst.action" target="_blank" class="ml14">京ICP备19004658号</a></p>            <p class="fz12_baidu">©1999-2019 北京创新乐知网络技术有限公司 </p>            </div>        </div>        <div class="allow-info-box">        <p><a href="https://csdnimg.cn/cdn/content-toolbar/csdn-ICP.png" target="_blank">经营性网站备案信息</a>        <em class="width126"><a href="http://www.cyberpolice.cn/" target="_blank"><span>网络110报警服务</span></a></em></p>        <p><a href="http://www.bjjubao.org/" target="_blank"><span>北京互联网违法和不良信息举报中心</span></a></p>        <p><a href="http://www.12377.cn/" target="_blank"><span>中国互联网举报中心</span></a><a href="https://download.csdn.net/index.php/tutelage/" target="_blank"><span style="margin-left:8px">家长监护</span></a><a href="https://blog.csdn.net/blogdevteam/article/details/90369522" target="_blank"><span style="margin-left:8px">版权申诉</span></a></p>        </div>        </div></div>
    	</div>
    </div>
    
    评论
    添加红包

    请填写红包祝福语或标题

    红包个数最小为10个

    红包金额最低5元

    当前余额3.43前往充值 >
    需支付:10.00
    成就一亿技术人!
    领取后你会自动成为博主和红包主的粉丝 规则
    hope_wisdom
    发出的红包
    实付
    使用余额支付
    点击重新获取
    扫码支付
    钱包余额 0

    抵扣说明:

    1.余额是钱包充值的虚拟货币,按照1:1的比例进行支付金额的抵扣。
    2.余额无法直接购买下载,可以购买VIP、付费专栏及课程。

    余额充值