blocks|key|1354556|text|这是实现您自己的评分对象[1]的惯例。它必须是正的，因为你可以创建一个不丢失的函数来计算一个自定义的正分数。这意味着，通过使用损失函数(对于得分对象)，您必须将负值。|type|unstyled|depth|inlineStyleRanges|entityRanges|offset|length|data|1354557|损失函数的范围是：(optimum)+[0.+...+%2B]+(e.g.+unequal+values+between+y+and+y')。例如，检查均方误差的公式，它总是肯定的：|style|CODE|1354558|​|1354559|📷|atomic|1354560|1354561|图片来源：求值.mean#均方误差|1354562|entityMap|0|LINK|mutability|MUTABLE|url|http://scikit-learn.org/stable/modules/model_evaluation.html#implementing-your-own-scoring-object|1|IMAGE|IMMUTABLE|imageUrl|https://developer.qcloudimg.com/http-save/yehe-900000/3a9486ac3bbf195bf4ef645f75a4e156.png|imageAlt|2|http://scikit-learn.org/stable/modules/model_evaluation.html#mean-squared-error^0|D|1|0|0|9|1N|0|0|0|1|1|0|0|5|C|2|0^^$0|@$1|2|3|4|5|6|7|15|8|@]|9|@$A|16|B|17|1|18]]|C|$]]|$1|D|3|E|5|6|7|19|8|@$A|1A|B|1B|F|G]]|9|@]|C|$]]|$1|H|3|I|5|6|7|1C|8|@]|9|@]|C|$]]|$1|J|3|K|5|L|7|1D|8|@]|9|@$A|1E|B|1F|1|1G]]|C|$]]|$1|M|3|I|5|6|7|1H|8|@]|9|@]|C|$]]|$1|N|3|O|5|6|7|1I|8|@]|9|@$A|1J|B|1K|1|1L]]|C|$]]|$1|P|3|-4|5|6|7|1M|8|@]|9|@]|C|$]]]|Q|$R|$5|S|T|U|C|$V|W]]|X|$5|Y|T|Z|C|$10|11|12|-4]]|13|$5|S|T|U|C|$V|14]]]]

It's a convention for implementing your own scoring object [<a href="http://scikit-learn.org/stable/modules/model_evaluation.html#implementing-your-own-scoring-object" rel="nofollow noreferrer">1</a>]. And it must be positive, because you can create a non-loss function to compute a custom positive score. That means that by using a loss function (for a score object) you have to the negative value.

The range of a loss function is: <code>(optimum) [0. ... +] (e.g. unequal values between y and y')</code>. For instance check the formula of the mean squared error, it's always positive:

<a href="https://i.stack.imgur.com/pGjVk.png" rel="nofollow noreferrer"><img src="https://i.stack.imgur.com/pGjVk.png" alt="mean squared error"></a>

Image source: <a href="http://scikit-learn.org/stable/modules/model_evaluation.html#mean-squared-error" rel="nofollow noreferrer">http://scikit-learn.org/stable/modules/model_evaluation.html#mean-squared-error</a>

blocks|key|82546|text|实际函数"mean_squared_error"没有任何关于负部分的内容。但是当您尝试'neg_mean_squared_error‘时实现的函数将返回一个被否定的分数版本。|type|unstyled|depth|inlineStyleRanges|offset|length|style|CODE|entityRanges|data|82547|请检查源代码如何在源代码中定义|82548|neg_mean_squared_error_scorer+=+make_scorer(mean_squared_error,
++++++++++++++++++++++++++++++++++++++++greater_is_better=False)|code-block|syntax|javascript|82549|观察如何将param+greater_is_better设置为False。|82550|现在，所有这些分数/损失都被用于其他各种事情，如cross_val_score、cross_val_predict、GridSearchCV等。例如，在“accuracy_score”或“F1_+score”的情况下，分数越高越好，但在发生损失(错误)时，分数越低越好。为了以相同的方式处理它们，它返回负值。|82551|因此，这个实用程序是为了以相同的方式处理分数和损失，而不更改特定损失或分数的源代码。|82552|所以，你没有错过任何东西。您只需要处理想要使用损失函数的场景。如果您只想计算mean_squared_error，则只能使用mean_squared_error。但是，如果您想使用它来调优您的模型，或者使用Scikit中的实用程序来优化cross_validate，请使用'neg_mean_squared_error'。|82553|也许添加一些细节，我会解释更多。|82554|entityMap|0|LINK|mutability|MUTABLE|url|https://github.com/scikit-learn/scikit-learn/blob/master/sklearn/metrics/_scorer.py#L629^0|4|K|0|9|3|0|0|0|B|H|V|5|0|0|0|1Q|I|3S|O|0|0^^$0|@$1|2|3|4|5|6|7|14|8|@$9|15|A|16|B|C]]|D|@]|E|$]]|$1|F|3|G|5|6|7|17|8|@]|D|@$9|18|A|19|1|1A]]|E|$]]|$1|H|3|I|5|J|7|1B|8|@]|D|@]|E|$K|L]]|$1|M|3|N|5|6|7|1C|8|@$9|1D|A|1E|B|C]|$9|1F|A|1G|B|C]]|D|@]|E|$]]|$1|O|3|P|5|6|7|1H|8|@]|D|@]|E|$]]|$1|Q|3|R|5|6|7|1I|8|@]|D|@]|E|$]]|$1|S|3|T|5|6|7|1J|8|@$9|1K|A|1L|B|C]|$9|1M|A|1N|B|C]]|D|@]|E|$]]|$1|U|3|V|5|6|7|1O|8|@]|D|@]|E|$]]|$1|W|3|-4|5|6|7|1P|8|@]|D|@]|E|$]]]|X|$Y|$5|Z|10|11|E|$12|13]]]]

The actual function <code>&quot;mean_squared_error&quot;</code> doesn't have anything about the negative part. But the function implemented when you try 'neg_mean_squared_error' will return a negated version of the score.
Please check the source code as to how its defined in <a href="https://github.com/scikit-learn/scikit-learn/blob/master/sklearn/metrics/_scorer.py#L629" rel="noreferrer">the source code</a>:
<pre><code>neg_mean_squared_error_scorer = make_scorer(mean_squared_error,
 greater_is_better=False)
</code></pre>
Observe how the param <code>greater_is_better</code> is set to <code>False</code>.
Now all these scores/losses are used in various other things like cross_val_score, cross_val_predict, GridSearchCV etc. For example, in cases of 'accuracy_score' or 'f1_score', the higher score is better, but in case of losses (errors), lower score is better. To handle them both in same way, it returns the negative.
So this utility is made for handling the scores and losses in same way without changing the source code for the specific loss or score.
So, you did not miss anything. You just need to take care of the scenario where you want to use the loss function. If you only want to calculate the mean_squared_error you can use <code>mean_squared_error</code> only. But if you want to use it to tune your models, or cross_validate using the utilities present in Scikit, use <code>'neg_mean_squared_error'</code>.
Maybe add some details about that and I will explain more.

blocks|key|82585|text|这正是我在我的代码中所要寻找的，我试图破译和澄清rmse报告，以理解我的数据。|type|unstyled|depth|inlineStyleRanges|entityRanges|data|82586|在我的例子中，我使用这种方法来计算rmse。我该怎么读这些报告？更高是更好还是相反？|82587|def+rmsle_cv(model):
++++kf+=+KFold(n_folds,+random_state=42).get_n_splits(train)
++++rmse=+np.sqrt(-cross_val_score(model,+train,+y_train,+scoring="neg_mean_squared_error",+cv+=+kf))
++++return(rmse)


def+rmsle(y,+y_pred):
++++return+np.sqrt(mean_squared_error(y,+y_pred))|code-block|syntax|javascript|82588|在我的例子中，我得到了这些结果|82589|Lasso+score(cv):+0.1176+(0.0068)
ElasticNet+score(cv):+0.1177+(0.0068)
Ridge(01):+0.1270+(0.0097)
Gradient+Boosting+score(cv):+0.1443+(0.0109)
BayRidge(01):+0.1239+(0.0079)
Kernel+Ridge+score(cv):+0.1210+(0.0068)
Xgboost+score(cv):+0.1177+(0.0060)
LGBM+score(cv):+0.1158+(0.0064)|82590|entityMap^0|0|0|0|0|0^^$0|@$1|2|3|4|5|6|7|O|8|@]|9|@]|A|$]]|$1|B|3|C|5|6|7|P|8|@]|9|@]|A|$]]|$1|D|3|E|5|F|7|Q|8|@]|9|@]|A|$G|H]]|$1|I|3|J|5|6|7|R|8|@]|9|@]|A|$]]|$1|K|3|L|5|F|7|S|8|@]|9|@]|A|$G|H]]|$1|M|3|-4|5|6|7|T|8|@]|9|@]|A|$]]]|N|$]]

This is exactly what I am looking for in my code that I am trying to decipher and clarify the rmse reports to make sense of my data.
in my case, I am using this approach to calculate the rmse. How should I read the reports? Is higher better or is it the opposite?
<pre><code>def rmsle_cv(model):
 kf = KFold(n_folds, random_state=42).get_n_splits(train)
 rmse= np.sqrt(-cross_val_score(model, train, y_train, scoring=&quot;neg_mean_squared_error&quot;, cv = kf))
 return(rmse)


def rmsle(y, y_pred):
 return np.sqrt(mean_squared_error(y, y_pred))
</code></pre>
In my case, I am getting these results
<pre><code>Lasso score(cv): 0.1176 (0.0068)
ElasticNet score(cv): 0.1177 (0.0068)
Ridge(01): 0.1270 (0.0097)
Gradient Boosting score(cv): 0.1443 (0.0109)
BayRidge(01): 0.1239 (0.0079)
Kernel Ridge score(cv): 0.1210 (0.0068)
Xgboost score(cv): 0.1177 (0.0060)
LGBM score(cv): 0.1158 (0.0064)
</code></pre>

blocks|key|679169|text|这是语义学上的问题。当你使用“损失”这一术语时，很大程度上意味着你想把它降到最低。然而，当你谈论“分数”时，它意味着你想要最大化它(例如，分类器的准确性)。因此，在对回归模型打分时，最大分数意味着最小损失，即最大负数损失。|type|unstyled|depth|inlineStyleRanges|entityRanges|data|679170|entityMap^0|0^^$0|@$1|2|3|4|5|6|7|D|8|@]|9|@]|A|$]]|$1|B|3|-4|5|6|7|E|8|@]|9|@]|A|$]]]|C|$]]

It's a semantics kind of issue. When you use the term &quot;loss&quot; it is heavily implied that you want to minimize it. However when you talk about &quot;score&quot; it's implied that you want to maximize it (e.g. accuracy of a classifier). Therefore, when scoring a regression model, a maximal score would mean a minimal loss, i.e. maximize the negated loss.

In general, the <code>mean_squared_error</code> is the smaller the better. 

When I am using the sklearn metrics package, it says in the document pages: <a href="http://scikit-learn.org/stable/modules/model_evaluation.html" rel="noreferrer">http://scikit-learn.org/stable/modules/model_evaluation.html</a>

<blockquote>
 All scorer objects follow the convention that higher return values are
 better than lower return values. Thus metrics which measure the
 distance between the model and the data, like
 metrics.mean_squared_error, are available as neg_mean_squared_error
 which return the negated value of the metric.
</blockquote>

and 
<a href="https://i.stack.imgur.com/7XYb7.png" rel="noreferrer"><img src="https://i.stack.imgur.com/7XYb7.png" alt="enter image description here"></a>

However, if I go to: <a href="http://scikit-learn.org/stable/modules/generated/sklearn.metrics.mean_squared_error.html#sklearn.metrics.mean_squared_error" rel="noreferrer">http://scikit-learn.org/stable/modules/generated/sklearn.metrics.mean_squared_error.html#sklearn.metrics.mean_squared_error</a>

It says it is the <code>Mean squared error regression loss</code>, didn't say it is negated.

And if I looked at the source code and checked the example there:<a href="https://github.com/scikit-learn/scikit-learn/blob/a24c8b46/sklearn/metrics/regression.py#L183" rel="noreferrer">https://github.com/scikit-learn/scikit-learn/blob/a24c8b46/sklearn/metrics/regression.py#L183</a> it is doing the normal <code>mean squared error</code>, i.e. the smaller the better.

So I am wondering if I missed anything about the negated part in the document. Thanks!

Is sklearn.metrics.mean_squared_error the larger the better (negated)?

翻译质量差，导致语言生硬或混乱。

没有提供实际的解决方法或示例。

解答不清晰，无法理解或解决问题。

页面排版不美观，阅读体验差。

文章

问答

视频

学习中心

腾讯云实验室

直播

竞赛

腾讯云代码分析专区

腾讯iOA零信任安全管理系统专区

腾讯云架构师技术同盟交流圈

腾讯云数据库专区

腾讯云顾问专区

腾讯云原生专区

腾讯混元专区

腾讯云TCE专区

腾讯云Lighthouse专区

腾讯云HAI专区

腾讯云Edgeone专区

腾讯云存储专区

腾讯云智能专区

腾讯轻联专区 

腾讯云开发专区

TAPD专区

腾讯轻量云游戏服专区

腾讯云最具价值专家

腾讯云架构师技术同盟

腾讯云创作之星

腾讯云开发者先锋 

腾讯云代码助手

CODING DevOps

Cloud Studio

SDK中心

API中心

命令行工具

一般来说，mean_squared_error越小越好。当我使用sklearn度量包时，文档页面中写着：所有scorer对象都遵循这样的惯例:返回值越高，返回值越低。因此，度量模型与数据之间的距离的度量(如metrics.mean_squared_error )作为neg_mean_squared_error可用，它返...

问sklearn.metrics.mean_squared_error越大越好(否定)吗？
EN

Stack Overflow用户

社区

活动

圈层

关于

腾讯云开发者

热门产品

热门推荐

更多推荐

问sklearn.metrics.mean_squared_error越大越好(否定)吗？EN