<?xml version="1.0" encoding="utf-8" ?><rss version="2.0"><channel><title><![CDATA[Linyi_DanielWu的博客]]></title><description><![CDATA[]]></description><link>https://blog.youkuaiyun.com/Linyi_DanielWu</link><language>zh-cn</language><generator>https://blog.youkuaiyun.com/</generator><copyright><![CDATA[Copyright &copy; Linyi_DanielWu]]></copyright><item><title><![CDATA[多分类对数损失（Multi-Class Log-Loss）代码]]></title><link>https://blog.youkuaiyun.com/Linyi_DanielWu/article/details/108278465</link><guid>https://blog.youkuaiyun.com/Linyi_DanielWu/article/details/108278465</guid><author>Linyi_DanielWu</author><pubDate>Fri, 28 Aug 2020 14:05:38 +0800</pubDate><description><![CDATA[def multiclass_logloss(actual, predicted, eps=1e-15):
    """Logarithmic Loss  Metric
    :param actual: 包含actual target classes的数组
    :param predicted: 分类预测结果矩阵, 每个类别都有一个概率
    """
    # Convert 'actual' to a binary array if it's not already:
    if len.]]></description><category></category></item><item><title><![CDATA[Bert 论文中文翻译]]></title><link>https://blog.youkuaiyun.com/Linyi_DanielWu/article/details/97391840</link><guid>https://blog.youkuaiyun.com/Linyi_DanielWu/article/details/97391840</guid><author>Linyi_DanielWu</author><pubDate>Fri, 26 Jul 2019 15:29:37 +0800</pubDate><description><![CDATA[BERT：预训练的深度双向 Transformer 语言模型

Jacob Devlin；Ming-Wei Chang；Kenton Lee；Kristina Toutanova Google AI Language {jacobdevlin,mingweichang,kentonl,kristout}@google.com

摘要

我们提出了一种新的称为 BERT 的语言表示模型，BERT 代...]]></description><category></category></item><item><title><![CDATA[json.decoder.JSONDecodeError: Unexpected UTF-8 BOM (decode using utf-8-sig)解决办法]]></title><link>https://blog.youkuaiyun.com/Linyi_DanielWu/article/details/97249767</link><guid>https://blog.youkuaiyun.com/Linyi_DanielWu/article/details/97249767</guid><author>Linyi_DanielWu</author><pubDate>Thu, 25 Jul 2019 11:48:22 +0800</pubDate><description><![CDATA[notepad++打开文件 -&gt; 格式 -&gt;以utf-8无bom模式编码 -&gt;保存
]]></description><category></category></item><item><title><![CDATA[BERT 论文 中文翻译]]></title><link>https://blog.youkuaiyun.com/Linyi_DanielWu/article/details/93751300</link><guid>https://blog.youkuaiyun.com/Linyi_DanielWu/article/details/93751300</guid><author>Linyi_DanielWu</author><pubDate>Wed, 26 Jun 2019 16:40:11 +0800</pubDate><description><![CDATA[11
]]></description><category></category></item><item><title><![CDATA[secureCRT中文乱码解决办法]]></title><link>https://blog.youkuaiyun.com/Linyi_DanielWu/article/details/93465811</link><guid>https://blog.youkuaiyun.com/Linyi_DanielWu/article/details/93465811</guid><author>Linyi_DanielWu</author><pubDate>Mon, 24 Jun 2019 10:17:45 +0800</pubDate><description><![CDATA[Options ——&gt;Session Options——&gt;Appearance——&gt;character encoding的值改为utf-8。


]]></description><category></category></item><item><title><![CDATA[XGBoost、LightGBM的详细对比介绍]]></title><link>https://blog.youkuaiyun.com/Linyi_DanielWu/article/details/90665834</link><guid>https://blog.youkuaiyun.com/Linyi_DanielWu/article/details/90665834</guid><author>Linyi_DanielWu</author><pubDate>Wed, 29 May 2019 10:06:46 +0800</pubDate><description><![CDATA[本文按照这些方法出现的先后顺序叙述。

GBDT

梯度提升树实在提升树的基础上发展而来的一种使用范围更广的方法，当处理回归问题时，提升树可以看作是梯度提升树的特例(分类问题时是不是特例？)。 因为提升树在构建树每一步的过程中都是去拟合上一步获得模型在训练集上的残差。后面我们将会介绍，这个残存正好是损失函数的梯度，对应于GBDT每一步要拟合的对象。

主要思想

在目标函数所在的函数空间中做梯度下...]]></description><category></category></item><item><title><![CDATA[2019中国高校计算机大赛——大数据挑战赛]]></title><link>https://blog.youkuaiyun.com/Linyi_DanielWu/article/details/90482404</link><guid>https://blog.youkuaiyun.com/Linyi_DanielWu/article/details/90482404</guid><author>Linyi_DanielWu</author><pubDate>Thu, 23 May 2019 16:06:39 +0800</pubDate><description><![CDATA[记录比赛中收集到的算法和资料



文本匹配（语义相似度/行为相关性）技术综述：https://blog.youkuaiyun.com/lipengcn/article/details/85313971
]]></description><category></category></item><item><title><![CDATA[python3报错‘ascii’ codec can’t encode characters in position 0-1: ordinal not in range(128) 解决办法]]></title><link>https://blog.youkuaiyun.com/Linyi_DanielWu/article/details/90411347</link><guid>https://blog.youkuaiyun.com/Linyi_DanielWu/article/details/90411347</guid><author>Linyi_DanielWu</author><pubDate>Tue, 21 May 2019 17:40:08 +0800</pubDate><description><![CDATA[python3

在windows上运行没问题，上传至linux服务器报错 ：UnicodeEncodeError: ‘ascii’ codec can’t encode characters in position 0-1: ordinal not in range(128)

原因:默认编码为ascii，无法表示中文

加入代码以下代码：默认编码改为utf-8，完美解决！


import s...]]></description><category></category></item><item><title><![CDATA[论文解读《Automatic Text Scoring Using Neural Networks》]]></title><link>https://blog.youkuaiyun.com/Linyi_DanielWu/article/details/90375072</link><guid>https://blog.youkuaiyun.com/Linyi_DanielWu/article/details/90375072</guid><author>Linyi_DanielWu</author><pubDate>Mon, 20 May 2019 15:03:20 +0800</pubDate><description><![CDATA[论文使用用C&amp;W Embedding及LSTM作为基础，提出了新的文本自动评分模型，取得了好的效果。

  具体如下：

  为什么要搞出一个SSWESAugmented C&amp;W model？

-C&amp;W Embeddings

  在NNLM之后，在CBOW和skip-gram之前，2008年Collobert和Weston 提出的C&amp;...]]></description><category></category></item><item><title><![CDATA[NLP文本评分、文章筛选]]></title><link>https://blog.youkuaiyun.com/Linyi_DanielWu/article/details/90290038</link><guid>https://blog.youkuaiyun.com/Linyi_DanielWu/article/details/90290038</guid><author>Linyi_DanielWu</author><pubDate>Fri, 17 May 2019 11:29:38 +0800</pubDate><description><![CDATA[最近要做一个新闻质量判断的项目，记录一下过程和收集到的资料，先占个坑，内容先冗杂在一起，后期再拆分

5.17

Automatic Text Scoring Using Neural Networks（使用神经网络的自动文本评分）：

-论文原文地址：https://arxiv.org/abs/1606.04289

-论文解读《Automatic Text Scoring Using Neur...]]></description><category></category></item><item><title><![CDATA[手动实现简单的 文本聚类算法 词袋模型+consin相似度+UPGMA聚类]]></title><link>https://blog.youkuaiyun.com/Linyi_DanielWu/article/details/88627744</link><guid>https://blog.youkuaiyun.com/Linyi_DanielWu/article/details/88627744</guid><author>Linyi_DanielWu</author><pubDate>Sun, 17 Mar 2019 21:42:09 +0800</pubDate><description><![CDATA[全手写文本聚类算法 词袋模型--&gt;consin相似度--&gt;UPGMA聚类


punctunation = [",", ":", ".", "'", "?", "/", "-", "+", "&amp;", "(", ")"]
sent = ["I love sky, I love sea.", "i love sky, I love sea.", "I want to go to ...]]></description><category></category></item></channel></rss>