python對邏輯回歸進行顯著性,在python中計算邏輯回歸

I tried to calculate logical regression. I have the data as csv file.

it looks like

node_id,second_major,gender,major_index,year,dorm,high_school,student_fac

0,0,2,257,2007,111,2849,1

1,0,2,271,2005,0,51195,2

2,0,2,269,2007,0,21462,1

3,269,1,245,2008,111,2597,1

..........................

This is my coding.

import pandas as pd

import statsmodels.api as sm

import pylab as pl

import numpy as np

df = pd.read_csv("Reed98.csv")

print df.describe()

dummy_ranks = pd.get_dummies(df['second_major'], prefix='second_major')

cols_to_keep = ['second_major', 'dorm', 'high_school']

data = df[cols_to_keep].join(dummy_ranks.ix[:, 'year':])

train_cols = data.columns[1:]

# Index([gre, gpa, prestige_2, prestige_3, prestige_4], dtype=object)

logit = sm.Logit(data['second_major'], data[train_cols])

result = logit.fit()

print result.summary()

When I run the coding in python I got an error:

Traceback (most recent call last):

File "D:\project\logisticregression.py", line 24, in

result = logit.fit()

File "c:\python26\lib\site-packages\statsmodels-0.5.0-py2.6- win32.egg\statsmodels\discrete\discrete_model.py", line 282, in fit

disp=disp, callback=callback, **kwargs)

File "c:\python26\lib\site-packages\statsmodels-0.5.0-py2.6- win32.egg\statsmodels\discrete\discrete_model.py", line 233, in fit

disp=disp, callback=callback, **kwargs)

File "c:\python26\lib\site-packages\statsmodels-0.5.0-py2.6- win32.egg\statsmodels\base\model.py", line 291, in fit

hess=hess)

File "c:\python26\lib\site-packages\statsmodels-0.5.0-py2.6-win32.egg\statsmodels\base\model.py", line 341, in _fit_mle_newton

newparams = oldparams - np.dot(np.linalg.inv(H),

File "C:\Python26\Lib\site-packages\numpy\linalg\linalg.py", line 445, in inv

return wrap(solve(a, identity(a.shape[0], dtype=a.dtype)))

File "C:\Python26\Lib\site-packages\numpy\linalg\linalg.py", line 328, in solve

raise LinAlgError('Singular matrix')

LinAlgError: Singular matrix

How to rewrite the code?

解決方案

There's nothing wrong with your code. My guess is that you have missing values in your data. Try a dropna or use missing='drop' to Logit. You might also check that the right hand side is full rank np.linalg.matrix_rank(data[train_cols].values)