Start Contacting machine learning

There was no basis, and a little bit would be recorded here.

The square distance of the model prediction value to the real value can be used as the basis for filtering the correct model.

def error (F,x,y) {

return sum ((f (x)-y) **2)

}

F (x) is the model prediction value and Y is the actual value. If the model predicts a value equal to the actual value (an extremely good effect) then the difference is 0. The square is to take positive (error has positive negative), equivalent to the meaning of absolute value.

In multiple models, the smaller the error indicates that the model is closer to the actual data, but only as a basis for the filtering model, and based on other conditions.

Installation Key points:

Keep only one copy of the installation file, or it will cause a compilation error to be randomly referenced.

Mac comes with and Python under/system/library/frameworks/python.framework, while downloaded packages are typically installed into

/library/frameworks/python.framework/, so should first sudo ln-s/system/library/frameworks/python.framework/library/ Frameworks/python.framework

Only one installation is guaranteed. If you delete/system/library/frameworks/python.framework and then put your/library/frameworks/python.framework link past, in addition to the 2.7,

Also need the previous old version, the system will be dependent on the old version.

Download dmg file after installation, configure Export Pythonpath=/system/library/frameworks/python.framework/versions/2.7/extras/lib/python

Download scikit-learn-0.14.1.tar.gz for Mac version

After extracting the scikit-learn-0.14.1, go to the directory directly sudo python setup.py install.

In Eclipse, add the Sklearn to forced builtins manually.