spreg.likratiotest¶

spreg.likratiotest(reg0, reg1)[source]¶

Likelihood ratio test statistic [Gre03]

Parameters:

reg0regression object: for constrained model (H0)
reg1regression object: for unconstrained model (H1)

Returns:

likratiodictionary: contains the statistic (likr), the degrees of freedom (df) and the p-value (pvalue)
likrfloat: likelihood ratio statistic
dfinteger: degrees of freedom
p-valuefloat: p-value

Examples

>>> import numpy as np
>>> import libpysal
>>> from libpysal import examples
>>> import scipy.stats as stats
>>> from spreg import ML_Lag, OLS
>>> from spreg import likratiotest

Use the baltim sample data set

>>> db = libpysal.io.open(examples.get_path("baltim.dbf"),'r')
>>> y_name = "PRICE"
>>> y = np.array(db.by_col(y_name)).T
>>> y.shape = (len(y),1)
>>> x_names = ["NROOM","NBATH","PATIO","FIREPL","AC","GAR","AGE","LOTSZ","SQFT"]
>>> x = np.array([db.by_col(var) for var in x_names]).T
>>> ww = libpysal.io.open(examples.get_path("baltim_q.gal"))
>>> w = ww.read()
>>> ww.close()
>>> w.transform = 'r'

OLS regression

>>> ols1 = OLS(y,x)

ML Lag regression

>>> mllag1 = ML_Lag(y,x,w)

>>> lr = likratiotest(ols1,mllag1)

>>> print("Likelihood Ratio Test: {0:.4f}       df: {1}        p-value: {2:.4f}".format(lr["likr"],lr["df"],lr["p-value"]))
Likelihood Ratio Test: 44.5721       df: 1        p-value: 0.0000