<?xml version="1.0" encoding="utf-8" ?><rss version="2.0"><channel><title><![CDATA[frogbar的博客]]></title><description><![CDATA[]]></description><link>https://blog.youkuaiyun.com/frogbar</link><language>zh-cn</language><generator>https://blog.youkuaiyun.com/</generator><copyright><![CDATA[Copyright &copy; frogbar]]></copyright><item><title><![CDATA[FM&FFM&DeepFFM]]></title><link>https://blog.youkuaiyun.com/frogbar/article/details/90054500</link><guid>https://blog.youkuaiyun.com/frogbar/article/details/90054500</guid><author>frogbar</author><pubDate>Fri, 10 May 2019 11:20:39 +0800</pubDate><description><![CDATA[https://blog.youkuaiyun.com/John_xyz/article/details/78933253



http://baijiahao.baidu.com/s?id=1579855367208283187&amp;wfr=spider&amp;for=pc

https://blog.youkuaiyun.com/g11d111/article/details/77430095

http:...]]></description><category></category></item><item><title><![CDATA[SQL时间戳与日期转换]]></title><link>https://blog.youkuaiyun.com/frogbar/article/details/80911933</link><guid>https://blog.youkuaiyun.com/frogbar/article/details/80911933</guid><author>frogbar</author><pubDate>Wed, 04 Jul 2018 14:36:32 +0800</pubDate><description><![CDATA[(1)日期转为时间戳UNIX_TIMESTAMP('2015-04-29','yyyy-MM-dd')(2)时间戳转为日期FROM_UNIXTIME('1430236800','yyyy-MM-dd')sql查询使用时可将两函数中第一个参数替换为相应的字段名称...]]></description><category></category></item><item><title><![CDATA[集成学习-Stacking]]></title><link>https://blog.youkuaiyun.com/frogbar/article/details/79528421</link><guid>https://blog.youkuaiyun.com/frogbar/article/details/79528421</guid><author>frogbar</author><pubDate>Mon, 12 Mar 2018 16:04:07 +0800</pubDate><description><![CDATA[转自  https://rasbt.github.io/mlxtend/user_guide/classifier/StackingClassifier/   (英文版)(感觉是原创)(推荐)http://blog.youkuaiyun.com/willduan1/article/details/73618677  (中文版)StackingClassifierAn ensemble-learning met...]]></description><category></category></item><item><title><![CDATA[决策树算法原理]]></title><link>https://blog.youkuaiyun.com/frogbar/article/details/79368976</link><guid>https://blog.youkuaiyun.com/frogbar/article/details/79368976</guid><author>frogbar</author><pubDate>Sun, 25 Feb 2018 15:12:09 +0800</pubDate><description><![CDATA[本文转自 https://www.cnblogs.com/pinard/p/6050306.html

http://www.cnblogs.com/pinard/p/6053344.html

 

ID3,C4.5见https://blog.youkuaiyun.com/lemon_tree12138/article/details/51837983

https://blog.youkuaiyun.com/lemo...]]></description><category></category></item><item><title><![CDATA[朴素贝叶斯算法]]></title><link>https://blog.youkuaiyun.com/frogbar/article/details/79235803</link><guid>https://blog.youkuaiyun.com/frogbar/article/details/79235803</guid><author>frogbar</author><pubDate>Fri, 02 Feb 2018 10:06:15 +0800</pubDate><description><![CDATA[转自
https://zhuanlan.zhihu.com/p/26262151
贝叶斯分类是一类分类算法的总称，这类算法均以贝叶斯定理为基础，故统称为贝叶斯分类。而朴素朴素贝叶斯分类是贝叶斯分类中最简单，也是常见的一种分类方法。这篇文章我尽可能用直白的话语总结一下我们学习会上讲到的朴素贝叶斯分类算法，希望有利于他人理解。

分类问题综述


对于分类问题，其实谁都不会陌生，日常生活]]></description><category></category></item><item><title><![CDATA[相关数据集地址]]></title><link>https://blog.youkuaiyun.com/frogbar/article/details/79230869</link><guid>https://blog.youkuaiyun.com/frogbar/article/details/79230869</guid><author>frogbar</author><pubDate>Thu, 01 Feb 2018 17:04:36 +0800</pubDate><description><![CDATA[spark  
sample_linear_regression_data.txt

http://code.taobao.org/p/bigdatas/diff/14/trunk/spark-1.0.1/mllib/data/sample_libsvm_data.txt]]></description><category></category></item><item><title><![CDATA[spark ML 与 MLlib 的区别]]></title><link>https://blog.youkuaiyun.com/frogbar/article/details/79228361</link><guid>https://blog.youkuaiyun.com/frogbar/article/details/79228361</guid><author>frogbar</author><pubDate>Thu, 01 Feb 2018 14:44:21 +0800</pubDate><description><![CDATA[https://www.zhihu.com/question/35225203
spark.mllib中的算法接口是基于RDDs的；spark.ml中的算法接口是基于DataFrames的。



技术角度上，面向的数据集类型不一样：ML的API是面向Dataset的（Dataframe是Dataset的子集，也就是Dataset[Row]）， mllib是面对RDD的。Dataset和]]></description><category></category></item><item><title><![CDATA[spark sql构建DataFrame从各种格式数据文件]]></title><link>https://blog.youkuaiyun.com/frogbar/article/details/79214498</link><guid>https://blog.youkuaiyun.com/frogbar/article/details/79214498</guid><author>frogbar</author><pubDate>Wed, 31 Jan 2018 10:56:28 +0800</pubDate><description><![CDATA[转自http://www.cnblogs.com/ywjy/p/7747482.html]]></description><category></category></item><item><title><![CDATA[idea shade 打包 & pom]]></title><link>https://blog.youkuaiyun.com/frogbar/article/details/79209718</link><guid>https://blog.youkuaiyun.com/frogbar/article/details/79209718</guid><author>frogbar</author><pubDate>Tue, 30 Jan 2018 19:30:42 +0800</pubDate><description><![CDATA[用于记录项目中的学习过程
pom配置如下部分中为shade插件,用于打成jar包,通过view->tool windows->Maven projects调出右边的Maven projects,依次点击plugins下的clean,然后lifecycle下的package则可完成打包.
xml version="1.0" encoding="UTF-8"?>
xmlns="http://mav]]></description><category></category></item><item><title><![CDATA[spark saveASTextFile]]></title><link>https://blog.youkuaiyun.com/frogbar/article/details/79174456</link><guid>https://blog.youkuaiyun.com/frogbar/article/details/79174456</guid><author>frogbar</author><pubDate>Fri, 26 Jan 2018 17:02:26 +0800</pubDate><description><![CDATA[转自 https://www.cnblogs.com/devilmaycry812839668/p/6922738.html
scala> val rd1 =sc.parallelize(Array(1 to 10000))
rd1: org.apache.spark.rdd.RDD[scala.collection.immutable.Range.Inclusive] = ParallelC]]></description><category></category></item><item><title><![CDATA[spark DataFrame 相关操作简记]]></title><link>https://blog.youkuaiyun.com/frogbar/article/details/79138713</link><guid>https://blog.youkuaiyun.com/frogbar/article/details/79138713</guid><author>frogbar</author><pubDate>Tue, 23 Jan 2018 13:41:33 +0800</pubDate><description><![CDATA[创建Dataset和Dataframe
最简单的方式就是使用spark.range方法来创建一个Dataset
val DS = spark.range(5, 100, 5)

创建DataFrame
val DF = spark.createDataFrame(List((&quot;Scala&quot;, 35), (&quot;Python&quot;, 30), (&quot;R&quot;, 15), (&quot;Java&quot;, 20)))  ]]></description><category></category></item><item><title><![CDATA[sparksession相关概念]]></title><link>https://blog.youkuaiyun.com/frogbar/article/details/79138705</link><guid>https://blog.youkuaiyun.com/frogbar/article/details/79138705</guid><author>frogbar</author><pubDate>Tue, 23 Jan 2018 13:40:20 +0800</pubDate><description><![CDATA[在2.0版本之前，与Spark交互之前必须先创建SparkConf和SparkContext，代码如下： 
val sparkConf = new SparkConf().setAppName("SparkSessionZipsExample").setMaster("local")  
val sc = new SparkContext(sparkConf).set("spark.some.]]></description><category></category></item><item><title><![CDATA[spark 部署方式]]></title><link>https://blog.youkuaiyun.com/frogbar/article/details/79138671</link><guid>https://blog.youkuaiyun.com/frogbar/article/details/79138671</guid><author>frogbar</author><pubDate>Tue, 23 Jan 2018 13:36:30 +0800</pubDate><description><![CDATA[spark部署方式
目前Apache Spark支持三种分布式部署方式，分别是standalone、spark on mesos和 spark on YARN


Standalone模式
即独立模式，自带完整的服务，可单独部署到一个集群中，无需依赖任何其他资源管理系统。


Spark On Mesos模式
官方推荐这种模式（当然，原因之一是血缘关系）.
1)粗粒度模式
1]]></description><category></category></item><item><title><![CDATA[spark-submit 相关参数]]></title><link>https://blog.youkuaiyun.com/frogbar/article/details/79138658</link><guid>https://blog.youkuaiyun.com/frogbar/article/details/79138658</guid><author>frogbar</author><pubDate>Tue, 23 Jan 2018 13:34:42 +0800</pubDate><description><![CDATA[spark-submit 相关参数


master url:
local: 使用1个worker线程在本地运行Spark程序
local[k]: 使用k个worker线程在本地运行Spark程序
local[*]: 使用所有剩余worker线程在本地运行Spark程序
spark://HOST:PORT: 连接到Spark Standalone集群,以便在该集群上运行Spark应用]]></description><category></category></item><item><title><![CDATA[spark-DataFrame操作]]></title><link>https://blog.youkuaiyun.com/frogbar/article/details/79081723</link><guid>https://blog.youkuaiyun.com/frogbar/article/details/79081723</guid><author>frogbar</author><pubDate>Wed, 17 Jan 2018 09:51:04 +0800</pubDate><description><![CDATA[http://blog.youkuaiyun.com/dabokele/article/details/52802150]]></description><category></category></item><item><title><![CDATA[python索引问题]]></title><link>https://blog.youkuaiyun.com/frogbar/article/details/78437599</link><guid>https://blog.youkuaiyun.com/frogbar/article/details/78437599</guid><author>frogbar</author><pubDate>Fri, 03 Nov 2017 17:17:24 +0800</pubDate><description><![CDATA[1.groupby后将索引变为数据框的列使用
2.某些csv文件读入后多列变为只有1列时使用
df1 = df.reset_index()]]></description><category></category></item><item><title><![CDATA[二分类相关评估指标（召回率、准确率，精确率，F度量，AUC和ROC）]]></title><link>https://blog.youkuaiyun.com/frogbar/article/details/78161546</link><guid>https://blog.youkuaiyun.com/frogbar/article/details/78161546</guid><author>frogbar</author><pubDate>Thu, 05 Oct 2017 15:40:53 +0800</pubDate><description><![CDATA[基础定义

通常在机器学习的二分类领域中，对模型的结果评估是必不可少的，本文主要总结了各个评估指标，对每个指标的定义，作用进行阐述。废话到此，直接上干货。

TP：True Positive  

FP：False Positive

TN：True Negtive

FN：False Negtive

以上四个定义是基础，Positive表示对样本作出的是正的判断，T表示判断正确，F表示判断错误...]]></description><category></category></item></channel></rss>