lawlite19 / machinelearning_python Goto Github PK

机器学习算法python实现

License: MIT License

Python 100.00%

machinelearning_python's Introduction

About me

Top Repositories

machinelearning_python's People

Contributors

Stargazers

Watchers

Forkers

pannyvan 0xqq shenxxxxxxxl loyalyonggang machinelearningorg 18786683795 hope0654 beartwocake hbishandb charlieyqin hardcao hao276843248 xxwudi508 stouduo piaoyats linecode greengithuber nickhell s654632396 pioneer12345 mary12956 x-nicolo qcloudsms stunum qxlei billy556 rainoftime dmairg fooszwt davidsxf marui1982 zhaojl1170 r00tak lovexj88 libennext chunleiml clove506 sjqzhang heblade luweidut othing2 sixleaves heiyixueren macleowh zhzgithub chenonly panlichen leemonkey greenary-john snzigod neverfinsh ds404 zephyrus9 xwttt vu1seek yefengjie jiangnancaizi matiji66 cailidong guanweidong xiaomaohoujiao2 wolfworld6 wqdsoft cuihee limm1997 weaponsandtools dongpingzhu midasc lixin-keyid yzcseu gokingstones lia-git yxy2829 shurenlee yohlee sutong-rm shanwf bodycoder101 mnlife novamartin wavinflaghxm foreverw ttklm20 brave731 520xiaopohai liangjunai zoery xclu chaoqunq wpf535236337 liu-hai-yang bruceanda clustersdata hsdong2012 bygreencn lxm909055383 datanewlook zhangxufang listeningandlistening kavoutsk

machinelearning_python's Issues

Trading model

How do I build a model to train my Trading EA Robot?

Please assist

关于反向传播的问题

在反向传播更新 Theta2_grad 时，delta3 是不是应该乘上 sigmoidGradient(z3) 之后再进行梯度更新呢

Font issue for MacOS

If I want to use OS to run the codes, the font address is not correct anymore.

font = FontProperties(fname=r"c:\windows\fonts\simsun.ttc", size=14) # 解决windows环境下画图汉字乱码问题

What should I do with this?

假负表达式错误，selectThreshold中fn = np.sum((cvPrecision == 1) & (yval == 0)).astype(float)

实际fn = np.sum((cvPrecision == 0) & (yval == 1)).astype(float)

关于nnGradient函数的一个问题

Theta1[:,0] = 0;
Theta2[:,0] = 0;
函数的这个地方，这样会改变nnParams的值，会直接将bias置0，
我试了下如果先copy的话，acc会达到99%+

请教如何建立这样的知识体系

楼主，你好！很希望能向你这样建立知识体系，想问一下有没有教程可以推荐，比如：如何搭建“readme”那样的目录，如何指引到对于的代码文件夹。（我是完全的github新手，问题可能有点傻，就是希望能建立你这样的知识体系，希望楼主能给一二指教。）非常感谢！！

Website development

pavannanimsd007.ezyro.com is website I developed

这里理论推导应该是负梯度方向啊，为什么这里用的是梯度方向，关键是结果还正确。

MachineLearning_Python/NeuralNetwok/NeuralNetwork.py

Line 165 in 7fc5684

    
           Theta2_grad = Theta2_grad+np.dot(np.transpose(delta3[i,:].reshape(1,-1)),a2[i,:].reshape(1,-1))

Hi sir I want to learn python machine learning for that I should want to learn python first or I want to start machine learning now

bug？

MachineLearning_Python/SVM/SVM_scikit-learn.py

Line 49 in 7eac77f

xp = np.linspace(np.min(X[:,0]),np.max(X[:,1]),100)

这一行构造等差数列写错？我觉得应该选取x轴的max和min吧

关于nn_params的问题

为什么要将nn_params reshape来reshape去呢，不能直接向optimize.fmin_cg函数中传入它本来的shape吗

LogisticRegression_scikit-learn.py

大神你好，非常感谢你写的这个python机器学习系列，受益良多！
今天在看LogisticRegression_scikit-learn.py时候，发现归一化那里，你使用了
# 归一化 scaler = StandardScaler() scaler.fit(x_train) x_train = scaler.fit_transform(x_train) x_test = scaler.fit_transform(x_test)
其中， scaler.fit(x_train)应该是不需要的吧?查了sklearn的文档，里面写道:
fit_transform(X[, y]) | Fit to data, then transform it.
而scaler.fit(x_train) 应该和scaler.transform(x_train)组合使用。
不知道理解的对不对，希望你有空的话可以帮忙看一下，感谢！
文档连接
http://scikit-learn.org/stable/modules/generated/sklearn.preprocessing.StandardScaler.html