<?xml version="1.0" encoding="utf-8" ?><rss version="2.0"><channel><title><![CDATA[Searching_Bird的专栏]]></title><description><![CDATA[]]></description><link>https://blog.youkuaiyun.com/Searching_Bird</link><language>zh-cn</language><generator>https://blog.youkuaiyun.com/</generator><copyright><![CDATA[Copyright &copy; Searching_Bird]]></copyright><item><title><![CDATA[Tensorflow 反卷积]]></title><link>https://blog.youkuaiyun.com/Searching_Bird/article/details/78376493</link><guid>https://blog.youkuaiyun.com/Searching_Bird/article/details/78376493</guid><author>Searching_Bird</author><pubDate>Sat, 28 Oct 2017 15:49:32 +0800</pubDate><description><![CDATA[Tenssorflow tf.nn.conv2d_tranpose() 反卷积]]></description><category></category></item><item><title><![CDATA[tensorflow 数据读取]]></title><link>https://blog.youkuaiyun.com/Searching_Bird/article/details/78331864</link><guid>https://blog.youkuaiyun.com/Searching_Bird/article/details/78331864</guid><author>Searching_Bird</author><pubDate>Tue, 24 Oct 2017 17:11:51 +0800</pubDate><description><![CDATA[tensorflow 数据读取的方式]]></description><category></category></item><item><title><![CDATA[tensorflow 变量创建，初始化，共享]]></title><link>https://blog.youkuaiyun.com/Searching_Bird/article/details/78295831</link><guid>https://blog.youkuaiyun.com/Searching_Bird/article/details/78295831</guid><author>Searching_Bird</author><pubDate>Fri, 20 Oct 2017 15:18:44 +0800</pubDate><description><![CDATA[tensorflow 变量创建，初始化，共享]]></description><category></category></item><item><title><![CDATA[tensorflow graph 中获取variable operation]]></title><link>https://blog.youkuaiyun.com/Searching_Bird/article/details/78283661</link><guid>https://blog.youkuaiyun.com/Searching_Bird/article/details/78283661</guid><author>Searching_Bird</author><pubDate>Thu, 19 Oct 2017 13:19:16 +0800</pubDate><description><![CDATA[tensorflow graph 中获取variable operation]]></description><category></category></item><item><title><![CDATA[python生成caffe binary proto文件]]></title><link>https://blog.youkuaiyun.com/Searching_Bird/article/details/78281115</link><guid>https://blog.youkuaiyun.com/Searching_Bird/article/details/78281115</guid><author>Searching_Bird</author><pubDate>Thu, 19 Oct 2017 10:14:27 +0800</pubDate><description><![CDATA[python 生成 caffe .binaryproto文件]]></description><category></category></item><item><title><![CDATA[tensorflow variable的保存和加载]]></title><link>https://blog.youkuaiyun.com/Searching_Bird/article/details/78274207</link><guid>https://blog.youkuaiyun.com/Searching_Bird/article/details/78274207</guid><author>Searching_Bird</author><pubDate>Wed, 18 Oct 2017 16:13:21 +0800</pubDate><description><![CDATA[tensorflow variable的保存和加载]]></description><category></category></item><item><title><![CDATA[Tensorflow Session graph Op 的理解]]></title><link>https://blog.youkuaiyun.com/Searching_Bird/article/details/78267510</link><guid>https://blog.youkuaiyun.com/Searching_Bird/article/details/78267510</guid><author>Searching_Bird</author><pubDate>Tue, 17 Oct 2017 22:07:19 +0800</pubDate><description><![CDATA[tensorflow graph session op tensor的理解]]></description><category></category></item><item><title><![CDATA[python读取LMDB中的图像]]></title><link>https://blog.youkuaiyun.com/Searching_Bird/article/details/78260598</link><guid>https://blog.youkuaiyun.com/Searching_Bird/article/details/78260598</guid><author>Searching_Bird</author><pubDate>Tue, 17 Oct 2017 15:01:10 +0800</pubDate><description><![CDATA[python解析LMDB中的数据。]]></description><category></category></item><item><title><![CDATA[caffe:图像数据 lmdb的写入]]></title><link>https://blog.youkuaiyun.com/Searching_Bird/article/details/51571407</link><guid>https://blog.youkuaiyun.com/Searching_Bird/article/details/51571407</guid><author>Searching_Bird</author><pubDate>Fri, 03 Jun 2016 01:36:56 +0800</pubDate><description><![CDATA[def image_dataset_to_lmdb(image_folder_path,lmdb_path,posAttributeList):        db  = lmdb.open(lmdb_path, map_size=2147483648)with db.begin(write=True) as db_txn:for parent, dirnames, filenames in os]]></description><category></category></item><item><title><![CDATA[caffe mean file：npy]]></title><link>https://blog.youkuaiyun.com/Searching_Bird/article/details/51571404</link><guid>https://blog.youkuaiyun.com/Searching_Bird/article/details/51571404</guid><author>Searching_Bird</author><pubDate>Fri, 03 Jun 2016 01:34:21 +0800</pubDate><description><![CDATA[blob = caffe.BlobProto()data = open(dbpath, "rb").read()blob.ParseFromString(data)nparray = blob.proto_to_array(blob)f = file("my_mean.npy","wb")np.save(f,nparray)]]></description><category></category></item><item><title><![CDATA[贝叶斯公式的一些理解]]></title><link>https://blog.youkuaiyun.com/Searching_Bird/article/details/46865363</link><guid>https://blog.youkuaiyun.com/Searching_Bird/article/details/46865363</guid><author>Searching_Bird</author><pubDate>Mon, 13 Jul 2015 17:28:30 +0800</pubDate><description><![CDATA[条件概率相关定理：





定理1

设A，B 是两个事件，且A不是不可能事件，则称


为在事件A发生的条件下，事件B发生的条件概率。一般地，


，且它满足以下三条件：

（1）非负性；（2）规范性；（3）可列可加性。

定理2

设E 为随机试验，Ω 为样本空间，A，B 为任意两个事件，设P(A)>0，称


为在“事件A 发生”的条件下事件B]]></description><category></category></item><item><title><![CDATA[分类器总结]]></title><link>https://blog.youkuaiyun.com/Searching_Bird/article/details/46860941</link><guid>https://blog.youkuaiyun.com/Searching_Bird/article/details/46860941</guid><author>Searching_Bird</author><pubDate>Mon, 13 Jul 2015 12:19:15 +0800</pubDate><description><![CDATA[]]></description><category></category></item><item><title><![CDATA[TF-IDF算法]]></title><link>https://blog.youkuaiyun.com/Searching_Bird/article/details/46797813</link><guid>https://blog.youkuaiyun.com/Searching_Bird/article/details/46797813</guid><author>Searching_Bird</author><pubDate>Wed, 08 Jul 2015 00:58:05 +0800</pubDate><description><![CDATA[TF-IDF及其算法


概念
     TF-IDF（term frequency–inverse document frequency）是一种用于资讯检索与资讯探勘的常用加权技术。TF-IDF是一种统计方法，用以评估一字词对于一个文件集或一个语料库中的其中一份文件的重要程度。字词的重要性随着它在文件中出现的次数成正比增加，但同时会随着它在语料库中出现的频率成反比下降。TF-IDF加权的]]></description><category></category></item><item><title><![CDATA[数据取对数运算的意义]]></title><link>https://blog.youkuaiyun.com/Searching_Bird/article/details/46797733</link><guid>https://blog.youkuaiyun.com/Searching_Bird/article/details/46797733</guid><author>Searching_Bird</author><pubDate>Wed, 08 Jul 2015 00:43:05 +0800</pubDate><description><![CDATA[平时在一些数据处理中，经常会把原始数据取对数后进一步处理。

之所以这样做是基于对数函数在其定义域内是单调增函数，取对数后不会改变数据的相对关系，取对数作用主要有：

1. 缩小数据的绝对数值，方便计算。

例如，每个数据项的值都很大，许多这样的值进行计算可能对超过常用数据类型的取值范围，这时取对数，就把数值缩小了，例如TF-IDF计算时，由于在大规模语料库中，很多词的频率是非常大的数]]></description><category></category></item><item><title><![CDATA[基于朴素贝叶斯到中文垃圾邮件分类器]]></title><link>https://blog.youkuaiyun.com/Searching_Bird/article/details/46783273</link><guid>https://blog.youkuaiyun.com/Searching_Bird/article/details/46783273</guid><author>Searching_Bird</author><pubDate>Tue, 07 Jul 2015 01:11:12 +0800</pubDate><description><![CDATA[简介：
   朴素 贝叶斯垃圾邮件分类器是在对邮件关键字进行统计分析到基础上利用贝叶斯公式进行分类到方法。相比其他模型，朴素贝叶斯模型到搭建比较简单，并且具有良好到分类效果。
   朴素真贝叶斯分类器通过对邮件关键字进行统计，然后使用贝叶斯推理来计算一封电子邮件是或不是垃圾邮件的概率。
贝叶斯定理：
     表示事件B已经发生的前提下，事件A发生的概率，叫做事件B发生下事件A的条件概率。]]></description><category></category></item><item><title><![CDATA[分类器的种类]]></title><link>https://blog.youkuaiyun.com/Searching_Bird/article/details/45999769</link><guid>https://blog.youkuaiyun.com/Searching_Bird/article/details/45999769</guid><author>Searching_Bird</author><pubDate>Tue, 26 May 2015 00:32:55 +0800</pubDate><description><![CDATA[分类器总体来说可以分成如下两类
分类器


统计分类器


回归分类器朴素贝叶斯贝叶斯网络


结构化分类器


基于规则的分类器基于距离的分类器神经网络]]></description><category></category></item><item><title><![CDATA[人工智能基本知识体系]]></title><link>https://blog.youkuaiyun.com/Searching_Bird/article/details/45772337</link><guid>https://blog.youkuaiyun.com/Searching_Bird/article/details/45772337</guid><author>Searching_Bird</author><pubDate>Sat, 16 May 2015 22:50:05 +0800</pubDate><description><![CDATA[AI概念：
人工智能=人造物（计算机）+智能（特殊化程序）
智能体（intelligent agent):自主决策
智能体的基本能力：自主性，学习进化，社会性（agent之间的协作交互）。
Part 1: 问题的求解


搜索技术：问题的答案可以通过搜索的方式得到则解决问题是可以采用搜索技术

无信息搜索技术（盲目搜索，例如广度优先，深度优先）启发式搜索:在启发函数的引导下进行搜]]></description><category></category></item></channel></rss>