<?xml version="1.0" encoding="utf-8" ?><rss version="2.0"><channel><title><![CDATA[MESSI_JAMES的博客]]></title><description><![CDATA[]]></description><link>https://blog.youkuaiyun.com/MESSI_JAMES</link><language>zh-cn</language><generator>https://blog.youkuaiyun.com/</generator><copyright><![CDATA[Copyright &copy; MESSI_JAMES]]></copyright><item><title><![CDATA[python知识互联]]></title><link>https://blog.youkuaiyun.com/MESSI_JAMES/article/details/134055957</link><guid>https://blog.youkuaiyun.com/MESSI_JAMES/article/details/134055957</guid><author>MESSI_JAMES</author><pubDate>Thu, 26 Oct 2023 14:42:46 +0800</pubDate><description><![CDATA[有关于机器学习或者python相关问题的可以通过邮箱13163378857.com发送给我，有时间回回复的，互相学习！]]></description><category></category></item><item><title><![CDATA[斐波那契实现]]></title><link>https://blog.youkuaiyun.com/MESSI_JAMES/article/details/115672327</link><guid>https://blog.youkuaiyun.com/MESSI_JAMES/article/details/115672327</guid><author>MESSI_JAMES</author><pubDate>Tue, 13 Apr 2021 16:42:41 +0800</pubDate><description><![CDATA[def F(x):
    if x&lt;2:
        return x
    else:
        return F(x-1)+F(x-2)
[F(i) for i in range(10)]


[0, 1, 1, 2, 3, 5, 8, 13, 21, 34]


]]></description><category></category></item><item><title><![CDATA[异常处理]]></title><link>https://blog.youkuaiyun.com/MESSI_JAMES/article/details/115671879</link><guid>https://blog.youkuaiyun.com/MESSI_JAMES/article/details/115671879</guid><author>MESSI_JAMES</author><pubDate>Tue, 13 Apr 2021 16:32:39 +0800</pubDate><description><![CDATA[我们常常想在没有触发异常的时候执⾏⼀些代码，这可以很轻松地通过⼀个else从句来达到。
有⼈也许问了：如果你只是想让⼀些代码在没有触发异常的情况下执⾏，为啥你不直接把代码放在try⾥⾯呢？
回答是，那样的话这段代码中的任意异常都还是会被try捕获，⽽你并不⼀定想要那样。
⼤多数⼈并不使⽤else从句，⽽且坦率地讲我⾃⼰也没有⼤范围使⽤。
这⾥是个例⼦：

try:
	print('I am sure no exception is going to occur!') 
except Exception: .]]></description><category></category></item><item><title><![CDATA[多项式拟合]]></title><link>https://blog.youkuaiyun.com/MESSI_JAMES/article/details/112276864</link><guid>https://blog.youkuaiyun.com/MESSI_JAMES/article/details/112276864</guid><author>MESSI_JAMES</author><pubDate>Wed, 06 Jan 2021 15:36:13 +0800</pubDate><description><![CDATA[import matplotlib.pyplot as plt
from scipy.optimize import curve_fit
import pandas as pd
import numpy as np

df=pd.read_excel('D:/CPO/BEV残值逻辑修改/TP超参数.xlsx')



x=df.TP_2.values
y=df.CS_2.values
z1 = np.polyfit(x, y, 7)
p1 = np.poly1d(z1)
list(p1)


# 画趋势图
]]></description><category></category></item><item><title><![CDATA[python获取mysql表头名称]]></title><link>https://blog.youkuaiyun.com/MESSI_JAMES/article/details/106381461</link><guid>https://blog.youkuaiyun.com/MESSI_JAMES/article/details/106381461</guid><author>MESSI_JAMES</author><pubDate>Wed, 27 May 2020 14:48:07 +0800</pubDate><description><![CDATA[import pymysql
import pandas as pd
import numpy as np
import datetime

def tp_data():
    #获取数据
    db=pymysql.connect(host=,port=3306,user=,password=)
    cursor=db.cursor()
    cursor.execute()
    data=cursor.fetchall()
    data_dict=[]
    for field in]]></description><category></category></item><item><title><![CDATA[特征工程之时间处理]]></title><link>https://blog.youkuaiyun.com/MESSI_JAMES/article/details/98767167</link><guid>https://blog.youkuaiyun.com/MESSI_JAMES/article/details/98767167</guid><author>MESSI_JAMES</author><pubDate>Wed, 07 Aug 2019 17:51:35 +0800</pubDate><description><![CDATA[def order_timestamp(x):
    result=datetime.datetime.strptime(x, '%Y-%m-%d %H:%M:%S.0')
    return result

def action_timestamp(timeNum):
    timeStamp = float(timeNum//1000)
    timeArray = time....]]></description><category></category></item><item><title><![CDATA[特征工程之标准化]]></title><link>https://blog.youkuaiyun.com/MESSI_JAMES/article/details/98763312</link><guid>https://blog.youkuaiyun.com/MESSI_JAMES/article/details/98763312</guid><author>MESSI_JAMES</author><pubDate>Wed, 07 Aug 2019 17:25:24 +0800</pubDate><description><![CDATA[数据和特征决定了机器学习算法的上限，而模型和算法只是不断逼近这个上限而已


数据标准化也就是统计数据的指数化。数据标准化处理主要包括数据同趋化处理和无量纲化处理两个方面：
数据同趋化处理主要解决不同性质数据问题，对不同性质指标直接加总不能正确反映不同作用力的综合结果，须先考虑改变逆指标数据性质，使所有指标对测评方案的作用力同趋化，再加总才能得出正确结果。
数据无量纲化处理主要解决数据的可比性。...]]></description><category></category></item><item><title><![CDATA[模型训练及调参]]></title><link>https://blog.youkuaiyun.com/MESSI_JAMES/article/details/98636236</link><guid>https://blog.youkuaiyun.com/MESSI_JAMES/article/details/98636236</guid><author>MESSI_JAMES</author><pubDate>Tue, 06 Aug 2019 17:54:43 +0800</pubDate><description><![CDATA[1.模型选择

根据二分类问题进行模型训练

from sklearn.linear_model import LogisticRegression
import pandas as pd
import numpy as np
from sklearn.neighbors import KNeighborsClassifier
from sklearn.linear_model import Lo...]]></description><category></category></item><item><title><![CDATA[Hadoop应用技术详解]]></title><link>https://blog.youkuaiyun.com/MESSI_JAMES/article/details/89326061</link><guid>https://blog.youkuaiyun.com/MESSI_JAMES/article/details/89326061</guid><author>MESSI_JAMES</author><pubDate>Tue, 16 Apr 2019 09:11:05 +0800</pubDate><description><![CDATA[下载地址
https://pan.baidu.com/s/1kTDWB6B
目录


]]></description><category></category></item><item><title><![CDATA[pycharm连接oracle中jdbc配置]]></title><link>https://blog.youkuaiyun.com/MESSI_JAMES/article/details/88062409</link><guid>https://blog.youkuaiyun.com/MESSI_JAMES/article/details/88062409</guid><author>MESSI_JAMES</author><pubDate>Fri, 01 Mar 2019 16:33:53 +0800</pubDate><description><![CDATA[1.配置方式：jdbc:oracle:thin:@//:1521/netgrid
2.配置方式：jdbc:oracle:thin:@:1521:netgrid
1中当@后边有双斜杠，端口号后边有单斜杠时，用服务名和实例名都可以链接成功，
2中@后边没有双斜杠，端口号后边是冒号而不是斜杠的时候只能用实例名，





...]]></description><category></category></item><item><title><![CDATA[python画小猪佩奇]]></title><link>https://blog.youkuaiyun.com/MESSI_JAMES/article/details/86551808</link><guid>https://blog.youkuaiyun.com/MESSI_JAMES/article/details/86551808</guid><author>MESSI_JAMES</author><pubDate>Sat, 19 Jan 2019 14:08:58 +0800</pubDate><description><![CDATA[https://github.com/Monster12138/-/blob/master/小猪佩奇.py
turtle
import turtle as t

t.pensize(4)
t.hideturtle()
t.colormode(255)
t.color((255,155,192),&amp;quot;pink&amp;quot;)
t.setup(840,500)
t.speed(10)

#鼻子
t.pu()
t.g...]]></description><category></category></item><item><title><![CDATA[AlexNet对MNIST分类]]></title><link>https://blog.youkuaiyun.com/MESSI_JAMES/article/details/85321418</link><guid>https://blog.youkuaiyun.com/MESSI_JAMES/article/details/85321418</guid><author>MESSI_JAMES</author><pubDate>Fri, 28 Dec 2018 14:22:56 +0800</pubDate><description><![CDATA[一.Alexnet介绍
https://blog.youkuaiyun.com/MESSI_JAMES/article/details/81384534#t8
二.过程介绍

一次完整的训练模型和评估模型的过程一般分为 3 个步骤：
1.加载数据，
2.定义网络模型，
3. 训练模型和评估模型。

三.代码实现
import tensorflow as tf
# 1.加载数据

# 输入数据
from ten...]]></description><category></category></item><item><title><![CDATA[LSTM对MNIST数据集做分类]]></title><link>https://blog.youkuaiyun.com/MESSI_JAMES/article/details/85095862</link><guid>https://blog.youkuaiyun.com/MESSI_JAMES/article/details/85095862</guid><author>MESSI_JAMES</author><pubDate>Wed, 19 Dec 2018 14:13:03 +0800</pubDate><description><![CDATA[https://morvanzhou.github.io/tutorials/machine-learning/tensorflow/5-08-RNN2/
1.设置 RNN 的参数

RNN 从每张图片的第一行像素读到最后一行, 然后再进行分类判断.

（1）导入 MNIST 数据并确定 RNN 的各种参数
下载地址: http://yann.lecun.com/exdb/mnist/

impo...]]></description><category></category></item><item><title><![CDATA[求列表的全子集]]></title><link>https://blog.youkuaiyun.com/MESSI_JAMES/article/details/84991920</link><guid>https://blog.youkuaiyun.com/MESSI_JAMES/article/details/84991920</guid><author>MESSI_JAMES</author><pubDate>Thu, 13 Dec 2018 17:17:31 +0800</pubDate><description><![CDATA[def getRealSubSet(fromList,toList):
    if(len(fromList) &amp;lt;= 1):
        return
    for id in range(len(fromList)):
        print('id',id)------------------------------------&amp;gt;打印看过程
        arr = ...]]></description><category></category></item><item><title><![CDATA[路径优化搜素算法]]></title><link>https://blog.youkuaiyun.com/MESSI_JAMES/article/details/84432586</link><guid>https://blog.youkuaiyun.com/MESSI_JAMES/article/details/84432586</guid><author>MESSI_JAMES</author><pubDate>Mon, 26 Nov 2018 11:42:19 +0800</pubDate><description><![CDATA[一.深度优先搜索算法(DFS)
1.算法介绍
https://zh.wikipedia.org/wiki/深度优先搜索

DFS（Depth-First-Search）是一种用于遍历或搜索树或图的算法。沿着树的深度遍历树的节点，尽可能深的搜索树的分支。当节点v的所在边都己被探寻过，搜索将回溯到发现节点v的那条边的起始节点。这一过程一直进行到已发现从源节点可达的所有节点为止。如果还存在未被发现的节点...]]></description><category></category></item><item><title><![CDATA[Tensorflow拟合函数：y=2x+3]]></title><link>https://blog.youkuaiyun.com/MESSI_JAMES/article/details/81837929</link><guid>https://blog.youkuaiyun.com/MESSI_JAMES/article/details/81837929</guid><author>MESSI_JAMES</author><pubDate>Sun, 19 Aug 2018 14:44:37 +0800</pubDate><description><![CDATA[
  本文代码在jupyter上实现 
  拟合一个一元一次函数：y=2x+3，具体要求如下： 
  1、使用Tensorflow框架 
  2、利用迭代方法，优化器不限 
  3、给出拟合的参数值 
  4、把拟合过程可视化


1.导入模块

import tensorflow as tf
import numpy as np
import matplotlib.pyplot as plt
%...]]></description><category></category></item><item><title><![CDATA[神经网路反向传播（BP）算法原理]]></title><link>https://blog.youkuaiyun.com/MESSI_JAMES/article/details/81749509</link><guid>https://blog.youkuaiyun.com/MESSI_JAMES/article/details/81749509</guid><author>MESSI_JAMES</author><pubDate>Thu, 16 Aug 2018 21:50:49 +0800</pubDate><description><![CDATA[一.BP算法简介


  BP算法的学习过程由正（前）向传播过程和反向传播过程组成。


1.正向传播


  将训练集数据输入到ANN的输入层，经过隐藏层，最后达到输出层并输出结果；




2.反向传播


  由于ANN的输出结果与实际结果有误差，则计算估计值与实际值之间的误差，并将该误差从输出层向隐藏层反向传播，直至传播到输入层。 
  在反向传播的过程中，根据误差调整各种参数的值；不断迭...]]></description><category></category></item><item><title><![CDATA[python基于神经网络实现人脸识别]]></title><link>https://blog.youkuaiyun.com/MESSI_JAMES/article/details/81704124</link><guid>https://blog.youkuaiyun.com/MESSI_JAMES/article/details/81704124</guid><author>MESSI_JAMES</author><pubDate>Wed, 15 Aug 2018 14:52:48 +0800</pubDate><description><![CDATA[一.项目简介



1.数据

需要数据QQ联系：1209028652


  数据集：总共数据集由两部分组成：他人脸图片集及我自己的部分图片 
  自己图片目录：face_recog/my_faces 
  他人图片目录：face_recog/other_faces 
  我的测试图片目录：face_recog/test_faces


2.人脸识别


  获取数据后，第一件事就对对图片进行处...]]></description><category></category></item><item><title><![CDATA[深度神经网络优化策略汇总]]></title><link>https://blog.youkuaiyun.com/MESSI_JAMES/article/details/81566106</link><guid>https://blog.youkuaiyun.com/MESSI_JAMES/article/details/81566106</guid><author>MESSI_JAMES</author><pubDate>Fri, 10 Aug 2018 16:46:57 +0800</pubDate><description><![CDATA[接下来介绍卷积神经网络的各种改进措施，其中经典网络的改进措施已经在前面各个网络中介绍。针对卷积神经网络的改进措施主要在以下几个方面：卷积层，池化层，激活函数，损失函数，网络结构，正则化技术等方面。优化算法对网络的训练至关重要，在这里我们单独列出来了。

1.卷积层


  卷积层的改进有以下几种：卷积核小型化，1x1 卷积，Network In Network，Inception机制，卷积分解（F...]]></description><category></category></item><item><title><![CDATA[多种类型RNN]]></title><link>https://blog.youkuaiyun.com/MESSI_JAMES/article/details/81542017</link><guid>https://blog.youkuaiyun.com/MESSI_JAMES/article/details/81542017</guid><author>MESSI_JAMES</author><pubDate>Thu, 09 Aug 2018 18:28:07 +0800</pubDate><description><![CDATA[1.RNN

https://zhuanlan.zhihu.com/p/36101196?utm_source=qq&amp;amp;amp;amp;amp;amp;amp;amp;amp;utm_medium=social&amp;amp;amp;amp;amp;amp;amp;amp;amp;utm_oi=761548970097917952


  （1）我们从基础的神经网络中知道，神经网络包含输入层、隐层、输出层，通过激活函数控制输出，层与层之间通过权值连接。激活函数是事先确定好的，那么神经网络模型]]></description><category></category></item></channel></rss>