libsvm

# libsvm

##  台湾大学林智仁教授主页    
LIBSVM -- A Library for Support Vector Machines
https://www.csie.ntu.edu.tw/~cjlin
svm_tutorial
https://www.csie.ntu.edu.tw/~piaip/svm/svm_tutorial.html
sample
https://www.csie.ntu.edu.tw/~cjlin/papers/guide/data/
explained
http://bytesizebio.net/2014/02/05/support-vector-machines-explained-well/

### libsvm动手实践
a. 简单应用  
./svm-train test.1  
./svm-scale test.1  
./svm-scale -l -1 -u 1 -s test.1.model.range test.1 > test.1.model.scale  
./svm-scale -s test.1.model.range test.1 > test.1.model.scale  
./svm-predict test.1 test.1.model test.1.predict  

b. 工具脚本应用  
检查数据  
python checkdata.py test.1  
优选参数Best c=8192.0, g=0.5 CV rate=97.1    
python grid.py test.1   
一步配置到位(其对grid.py、svm-train、svm-scale和svm-predict都进行了调用)  
python easy.py test.1  

c. 简单实例化  
数据格式转换  
python libsvm2csv.py test.1 test.11 4  
python csv2libsvm.py test.11 test.111  
test.111应该还原成test.1

 

### demo

1. https://github.com/clysto/libsvm-demo

 

2. 性别预测小例子

https://blog.csdn.net/zilongreco/article/details/41390385?utm_medium=distribute.pc_relevant_download.none-task-blog-baidujs-1.nonecase&depth_1-utm_source=distribute.pc_relevant_download.none-task-blog-baidujs-1.nonecase

在linux下整理了个Makefile工程

https://files.cnblogs.com/files/dong1/svmregress.tar.gz

 

3. svm.h里的几个接口都很容易用
svm_train
svm_save_model
svm_load_model
svm_predict
实际应用时直接加载model, 然后predict就行, 训练模型才需要配置参数，也可以不配置，默认也行. 

 

4. libsvm simplest demo !

https://github.com/dong2/2libsvm 

 

5. 加速度传感器检测步行，跑步，开车的示例
http://cs229.stanford.edu/proj2008/NhamSiangliulueYeung-PredictingModeOfTransportFromIphoneAccelerometerData.pdf

 

6. python-sklearn版本的Support-Vector-Machine

https://github.com/mahesh147/Support-Vector-Machine 安装numpy scipy pandas matplotlib```

sudo apt-get update
sudo apt-get install python-numpy
sudo apt-get install python-scipy
sudo apt-get install python-pandas
sudo apt-get install python-matplotlib
sudo apt-get install python-sklearn

sudo apt install python-pip
pip install --upgrade pip
pip install --user --upgrade scikit-learn --timeout=500

https://blog.csdn.net/yato0514/article/details/78754551
```

7. 另外一个python-sklearn

Support-Vector-Machine sklearn.svm
https://github.com/AlexanderFabisch/svm

 

### 更多资源(其实看官网就足够了)

https://github.com/xzbuck/kaggle/tree/dev/static/images/competitions/getting-started/digit-recognizer/svm  
http://www.jnlp.org/lab/graduates/sannomiya/libsvm  
入门指南翻译  
https://blog.csdn.net/sinat_25857925/article/details/70052532  
应用手册翻译  
https://blog.csdn.net/pangpang1239/article/details/7483895  
libsvm使用方法  
https://blog.csdn.net/yushupan/article/details/78998128  
测试数据  
https://www.csie.ntu.edu.tw/~cjlin/papers/guide/data/  
网友blog  
https://www.cnblogs.com/zhizhan/tag/svm/  
应用示例  
NhamSiangliulueYeung-PredictingModeOfTransportFromIphoneAccelerometerData.pdf  
https://www.baidu.com/link?url=WVMH-GPPwkBHcykH7qupVhxY_lKtMVTMzXoHKW1njvAmUOP6sMCZeoCIo37VzdhbR6HyAcag-Y2SMUbzb0JrBVU-hW7SwhBOJ0egYqLPC9xtKdx1TrzKPAj-B4lUoJW-jBseLm3elkSWxbPGiz7FzgK9pZghUfiuBtim-RtpZHa&wd=&eqid=cf5ced2f00021277000000066005595e  
libsvm回归  
http://blog.sina.com.cn/s/blog_5980835e0100drwx.html  
phraug/phraug2数据转换  
https://github.com/zygmuntz/phraug  
https://github.com/zygmuntz/phraug2