<?xml version="1.0" encoding="utf-8" ?><rss version="2.0"><channel><title><![CDATA[hyhpyx的博客]]></title><description><![CDATA[]]></description><link>https://blog.youkuaiyun.com/hyhpyx</link><language>zh-cn</language><generator>https://blog.youkuaiyun.com/</generator><copyright><![CDATA[Copyright &copy; hyhpyx]]></copyright><item><title><![CDATA[BiLSTM-CRF计算细节]]></title><link>https://blog.youkuaiyun.com/hyhpyx/article/details/122146667</link><guid>https://blog.youkuaiyun.com/hyhpyx/article/details/122146667</guid><author>hyhpyx</author><pubDate>Sun, 26 Dec 2021 13:34:08 +0800</pubDate><description><![CDATA[1、计算所有路径得分的对数指数和
  def _forward_alg(self, feats):
        init_alphas = torch.full((1, self.tagset_size), -10000.).cuda()
        init_alphas[0][self.tag_to_ix[START_TAG]] = 0.
        #注释1
        forward_var = init_alphas
        #注释2
        for feat in]]></description><category></category></item><item><title><![CDATA[从HMM到BiLSTM-CRF过程详解（1）]]></title><link>https://blog.youkuaiyun.com/hyhpyx/article/details/121889887</link><guid>https://blog.youkuaiyun.com/hyhpyx/article/details/121889887</guid><author>hyhpyx</author><pubDate>Mon, 13 Dec 2021 15:14:18 +0800</pubDate><description><![CDATA[背景知识
本小节主要对推导过程中存在的一些背景知识进行简要介绍，大家可以按照自己的情况进行阅读，有错误之处还请指出，谢谢啦~
1.1 逻辑回归与最大熵模型
1.1.1多分类逻辑回归模型与softmax函数
1.1.1.1.1 多分类逻辑回归函数
对于二分类逻辑回归函数想必大家肯定都很熟悉啦，我们在此简单回顾一下!逻辑回归模型如下：
P(Y=1∣X)=ewP(Y=1|X) = \frac{e^{w}}{}P(Y=1∣X)=ew​

...]]></description><category></category></item><item><title><![CDATA[机器学习第一话：逻辑回归（使用梯度下降法求解）]]></title><link>https://blog.youkuaiyun.com/hyhpyx/article/details/115096800</link><guid>https://blog.youkuaiyun.com/hyhpyx/article/details/115096800</guid><author>hyhpyx</author><pubDate>Mon, 22 Mar 2021 20:44:57 +0800</pubDate><description><![CDATA[逻辑回归
The data
我们将建立一个逻辑回归模型来预测一个学生是否被大学录取。假设你是一个大学系的管理员，你想根据两次考试的结果来决定每个申请人的录取机会。你有以前的申请人的历史数据，你可以用它作为逻辑回归的训练集。对于每一个培训例子，你有两个考试的申请人的分数和录取决定。为了做到这一点，我们将建立一个分类模型，根据考试成绩估计入学概率。
#三大件
import numpy as np
import pandas as pd
import matplotlib.pyplot as plt
%matpl]]></description><category></category></item><item><title><![CDATA[机器学习第一话：初识逻辑回归]]></title><link>https://blog.youkuaiyun.com/hyhpyx/article/details/115079600</link><guid>https://blog.youkuaiyun.com/hyhpyx/article/details/115079600</guid><author>hyhpyx</author><pubDate>Mon, 22 Mar 2021 20:15:34 +0800</pubDate><description><![CDATA[逻辑回归（Logistic regression）
什么是逻辑回归？
在我们以前的学习中，回归通常是指利用给定的样本信息X(x1,x2,...,xn)X(x_1,x2,...,x_n)X(x1​,x2,...,xn​)建立一个回归表达式，例如:
y=β0x1+β1x2+...+βnxxy = \beta_0x1+\beta1x2+...+\beta_nx_xy=β0​x1+β1x2+...+βn​xx​
最中我们得到是一个值y^\hat yy^​。对于预测某个样本是正例还是负例这种普通回归就不太适用，因为我]]></description><category></category></item><item><title><![CDATA[python 爬虫——快递物流查询]]></title><link>https://blog.youkuaiyun.com/hyhpyx/article/details/108945880</link><guid>https://blog.youkuaiyun.com/hyhpyx/article/details/108945880</guid><author>hyhpyx</author><pubDate>Wed, 07 Oct 2020 00:37:16 +0800</pubDate><description><![CDATA[import requests
import ast
num = input('请输入运单号：')
url = 'https://biz.trace.ickd.cn/yuantong/'+str(num)
response = requests.get(url = url,headers = headers)
response.encoding = response.apparent_encoding
page_txt = response.text
page_dict = ast.literal_eval]]></description><category></category></item><item><title><![CDATA[超级鹰使用流程]]></title><link>https://blog.youkuaiyun.com/hyhpyx/article/details/108945827</link><guid>https://blog.youkuaiyun.com/hyhpyx/article/details/108945827</guid><author>hyhpyx</author><pubDate>Wed, 07 Oct 2020 00:28:28 +0800</pubDate><description><![CDATA[超级鹰代码
超级鹰流程
1、首页 -&gt; 软件ID -&gt; 生成一个软件ID（并记住）
2、开发文档模块 -&gt; 下载python文档
import requests
from hashlib import md5

class Chaojiying_Client(object):

    def __init__(self, username, password, soft_id):
        self.username = username
        # 此处是经过修改后的代码]]></description><category></category></item><item><title><![CDATA[python环境下json文件的保存与读取]]></title><link>https://blog.youkuaiyun.com/hyhpyx/article/details/107875681</link><guid>https://blog.youkuaiyun.com/hyhpyx/article/details/107875681</guid><author>hyhpyx</author><pubDate>Sat, 08 Aug 2020 10:19:19 +0800</pubDate><description><![CDATA[保存数据
使用json.dump(需要被保存的内容,用于存储数据的文件对象)保存json文件
例：
例：

import json
numbers = [2,3,5,7,9,11]
with open(r'C:\Users\胡银洪\Desktop\jd.json','w') as f:
json.dump(numbers,f)


读取数据
使用json.load(f)读取文件
例：
with open(r'C:\Users\胡银洪\Desktop\jd.json') as f:
    numbers =]]></description><category></category></item><item><title><![CDATA[在jupyter notebook下python使用自定义模块]]></title><link>https://blog.youkuaiyun.com/hyhpyx/article/details/107854583</link><guid>https://blog.youkuaiyun.com/hyhpyx/article/details/107854583</guid><author>hyhpyx</author><pubDate>Fri, 07 Aug 2020 09:09:12 +0800</pubDate><description><![CDATA[前言：当引用模块时，编译器会先在当前目录（也就是所运行的文件所处的目录），接着去sys.path、Python的安装目录去寻找你引用的模块，如果没有的话，就会报错。注意寻找的先后顺序。
第一步：建立模块
所谓模块其实就是将多个函数封装为扩展名为.py的文件。
在jupyter创建.py文件的方法为：
1、先写出所需使用的函数。
例如：
def test1():
print(‘这是test1’)
def test2():
print(‘这是test2’)
2、将该.ipynb保存为.py文件。具体方法为：
点]]></description><category></category></item><item><title><![CDATA[matplotlib各种图的绘制]]></title><link>https://blog.youkuaiyun.com/hyhpyx/article/details/107366814</link><guid>https://blog.youkuaiyun.com/hyhpyx/article/details/107366814</guid><author>hyhpyx</author><pubDate>Wed, 15 Jul 2020 18:04:55 +0800</pubDate><description><![CDATA[折线图
plt.plot(x,y)

散点图
plt.scatter(x,y)

条形图
from matplotlib import pyplot as plt
from matplotlib import font_manager
my_font = font_manager.FontProperties(fname = r'C:\Windows\Fonts\simhei.ttf')
a = ['政治','英语二','数学三','统计学']
b = [82,78,122,120]
plt.figure(]]></description><category></category></item><item><title><![CDATA[python类的入门]]></title><link>https://blog.youkuaiyun.com/hyhpyx/article/details/107134697</link><guid>https://blog.youkuaiyun.com/hyhpyx/article/details/107134697</guid><author>hyhpyx</author><pubDate>Sun, 05 Jul 2020 09:52:00 +0800</pubDate><description><![CDATA[类的定义
class Person:   #名称首字母大写
    def setage(self,age):
        self.age = age
    def getage(self):
        print(self.age)
        

类的使用
#使用类的时候必须要实例化，而实例化之后的类的称为对象
p = Person()  ##这里p就是Person这个类的具体的实例对象,即self为p
p.setage(18)
p.getage()

魔法方法__init__
__i]]></description><category></category></item><item><title><![CDATA[常用sklearn库数据集调用方法]]></title><link>https://blog.youkuaiyun.com/hyhpyx/article/details/107080953</link><guid>https://blog.youkuaiyun.com/hyhpyx/article/details/107080953</guid><author>hyhpyx</author><pubDate>Thu, 02 Jul 2020 11:44:05 +0800</pubDate><description><![CDATA[所需第三方库
import numpy
import scipy
import matplotlib
import sklearn

boston房价
#导入boston房价数据集
from sklearn.datasets import load_boston  
#将数据集赋给boston
boston = load_boston()
#打印数据维度
print(boston.data.shape)  

#导入boston房价数据集
from sklearn.datasets import load_]]></description><category></category></item><item><title><![CDATA[初识pandas]]></title><link>https://blog.youkuaiyun.com/hyhpyx/article/details/106961002</link><guid>https://blog.youkuaiyun.com/hyhpyx/article/details/106961002</guid><author>hyhpyx</author><pubDate>Thu, 25 Jun 2020 18:38:17 +0800</pubDate><description><![CDATA[import pandas as pd
from IPython.display import display
data = {'name':['taiyo','anna','peter','hana'],
        'location':['new york','paris','berlin','london'],
        'age':[24,55,66,54]
    
}
data_pandas = pd.DataFrame(data)
#IPython.display可以在jupyte]]></description><category></category></item><item><title><![CDATA[python matplotlib库详解]]></title><link>https://blog.youkuaiyun.com/hyhpyx/article/details/106960738</link><guid>https://blog.youkuaiyun.com/hyhpyx/article/details/106960738</guid><author>hyhpyx</author><pubDate>Thu, 25 Jun 2020 18:07:30 +0800</pubDate><description><![CDATA[%matplotlib inline
#这是魔法函数，在jupyter notebook代替plt.show()
import matplotlib.pyplot as plt
# 在-10和10之间生成一个数列，共100个数
x = np.linspace(-10,10,100)
#用正弦函数创建第二个函数
y = np.sin(x)
#plot函数绘制一个数组关于另一个数组的折线图
plt.plot(x,y,marker = 'x')  



...]]></description><category></category></item><item><title><![CDATA[初识numpy,scipy]]></title><link>https://blog.youkuaiyun.com/hyhpyx/article/details/106960535</link><guid>https://blog.youkuaiyun.com/hyhpyx/article/details/106960535</guid><author>hyhpyx</author><pubDate>Thu, 25 Jun 2020 17:49:53 +0800</pubDate><description><![CDATA[from scipy import sparse
import numpy as np


x = np.array([[1,2,3],[4,5,6]])  #利用numpy创建一个数组
eye = np.eye(4) #创建一个4维单位向量
print('x:\n{}'.format(x))
print('eye:\n{}'.format(eye))


##稀疏矩阵是指0元素的个数远远小于非0元素
#可以通过numpy数组创建scipy稀疏矩阵，即只保留非0元素，并给出其位置，减少所用内存
sparse]]></description><category></category></item><item><title><![CDATA[python 文件操作及OSError: [Errno 22]解决]]></title><link>https://blog.youkuaiyun.com/hyhpyx/article/details/106623727</link><guid>https://blog.youkuaiyun.com/hyhpyx/article/details/106623727</guid><author>hyhpyx</author><pubDate>Mon, 08 Jun 2020 20:10:22 +0800</pubDate><description><![CDATA[打开文件及关闭文件
法一：使用raw string，在前面加上r
f = open(r'C:\Users\胡银洪\Desktop\picture\test.txt','w')
# 若在该路径下无该文件，w模式（写模式下）则会自动创建该名称文件
f.close()

法二：加入转义字符\
f = open('C:\\Users\\胡银洪\\Desktop\\picture\\test.txt','w')

法三：利用/代替\
f = open('C:/Users/胡银洪/Desktop/picture/tes]]></description><category></category></item><item><title><![CDATA[python 常用快捷键（更新中）]]></title><link>https://blog.youkuaiyun.com/hyhpyx/article/details/106624520</link><guid>https://blog.youkuaiyun.com/hyhpyx/article/details/106624520</guid><author>hyhpyx</author><pubDate>Mon, 08 Jun 2020 17:22:09 +0800</pubDate><description><![CDATA[ctrl+f ：打开查询框
Windows+shift+s:截图
ctrl+/:多行注释
shift+enter:运行代码

]]></description><category></category></item><item><title><![CDATA[pyhton 字符串]]></title><link>https://blog.youkuaiyun.com/hyhpyx/article/details/106619693</link><guid>https://blog.youkuaiyun.com/hyhpyx/article/details/106619693</guid><author>hyhpyx</author><pubDate>Mon, 08 Jun 2020 15:12:43 +0800</pubDate><description><![CDATA[字符串含有引号的情况
若字符串中含有单引号，则外层使用双引号
s = "I'm a student"
print(s)

若不使用双引号，可以通过\方式
s = 'I\'m a student'
print(s)

即通过\进行转义

]]></description><category></category></item><item><title><![CDATA[pyhton  break&continue]]></title><link>https://blog.youkuaiyun.com/hyhpyx/article/details/106615452</link><guid>https://blog.youkuaiyun.com/hyhpyx/article/details/106615452</guid><author>hyhpyx</author><pubDate>Mon, 08 Jun 2020 11:44:44 +0800</pubDate><description><![CDATA[break
从此处直接停止最内层循环
i = 0
while i&lt;10:
    i = i+1
    print('-'*30)
    if i==5:
        break  #从此处直接停止最内层循环
    print(i)

结果

为了检查break是否只是停止最内层循环，可设置两个循环，进行检验
j = 0
while j &lt; 3:
    j = j + 1
    i = 0
    while i &lt; 10:
        i = i+1
        p]]></description><category></category></item><item><title><![CDATA[python ttf字体文件]]></title><link>https://blog.youkuaiyun.com/hyhpyx/article/details/106600330</link><guid>https://blog.youkuaiyun.com/hyhpyx/article/details/106600330</guid><author>hyhpyx</author><pubDate>Mon, 08 Jun 2020 10:45:46 +0800</pubDate><description><![CDATA[发邮件至2532524866@qq.com，本人免费发送

]]></description><category></category></item><item><title><![CDATA[贪婪匹配&最小匹配]]></title><link>https://blog.youkuaiyun.com/hyhpyx/article/details/106555267</link><guid>https://blog.youkuaiyun.com/hyhpyx/article/details/106555267</guid><author>hyhpyx</author><pubDate>Sun, 07 Jun 2020 12:53:49 +0800</pubDate><description><![CDATA[贪婪匹配
m = re.search(r'PY.*N','PYGDHFNCDWJNFHJNFN')  #以PY开头，以N结尾,默认返回贪婪匹配，即返回最长的匹配
m.group(0)

最小匹配
m = re.search(r'PY.*?N','PYGDHFNCDWJNFHJNFN')  #最小匹配
m.group(0)

最小匹配操作符


...]]></description><category></category></item></channel></rss>