zhpmatrix

some codes are so interesting,so I just push them in the github,and share them with our friends.

zhpmatrix

#include <stdio.h>
int main(int argc,char* argv[]){
int i = 0;/* counter */
for(i = 0;i < 10000;i++){
   printf("just try more...\n");
}
return 0;
}

zhpmatrix

in most of the  times,I think you must think more,then just do it!that's what I really need...come on man...

zhpmatrix

Try your best!

zhpmatrix

在学习Scipy的leastsq扩展的时候，发现在中文前会用到'u'标签，这是干什么用的？

例如下述两行代码：

---

print u'真实参数',[A,k,theta]

pl.plot(x,y1,label = u'噪声数据')

为什么要加入u标签呢？否则中文报错！

zhpmatrix

scipy.sparse.coo_matrix的坑

---

上代码：

---

from scipy import sparse as sp
import numpy as np

zhpmatrix

1.when you know that your measurements of X are uncertain,

or when you don’t want to focus on the errors of one variable over another.

2.the history of the ODR:

Orthogonal Distance Regression (ODR) is a method that can do this (orthogonal in this context means perpendicular –

so it calculates errors perpendicular to the line, rather than just ‘vertically’).

zhpmatrix

下述TODO摘自知友（匿名）

---

*1.*比如ROC曲线的计算，自己写的ROC脚本就是很慢，但是sklearn里就是特别的快。对比之后发现，sklearn会非常善用numpy的一些特性，而不是用list去遍历。

*2.*在sklearn里，是直接使用c写了kdtree来实现KNN(Calling c/c++ from python ?)

*3.*虽然不少常用算法，比如decision tree用了cython，svm直接wrap了libsvm，但后期的代码基本原则都是优先Python实现，只有Python实在太慢的时候才会考虑用cython加速。

zhpmatrix

plt.xlabel("$x_1$")

plt.ylabel("$x_2$")

import numpy as np
import matplotlib.pyplot as plt

from sklearn.decomposition import PCA, KernelPCA
from sklearn.datasets import make_circles

zhpmatrix

'''载入训练数据'''

iris = load_iris()

X,y = iris.data,iris.target

'''打乱原始训练数据'''

'''获取行索引'''