Maximum Likelihood Estimation (Generic models)

This tutorial explains how to quickly implement new maximum likelihood models in statsmodels. We give two examples:

  1. Probit model for binary dependent variables
  2. Negative binomial model for count data

The GenericLikelihoodModel class eases the process by providing tools such as automatic numeric differentiation and a unified interface to scipy optimization functions. Using statsmodels, users can fit new MLE models simply by "plugging-in" a log-likelihood function.

Example 1: Probit model

In [1]:
from __future__ import print_function
import numpy as np
from scipy import stats
import statsmodels.api as sm
from statsmodels.base.model import GenericLikelihoodModel

The Spector dataset is distributed with statsmodels. You can access a vector of values for the dependent variable (endog) and a matrix of regressors (exog) like this:

In [2]:
data = sm.datasets.spector.load_pandas()
exog = data.exog
endog = data.endog
print(sm.datasets.spector.NOTE)
print(data.exog.head())
::

    Number of Observations - 32

    Number of Variables - 4

    Variable name definitions::

        Grade - binary variable indicating whether or not a student's grade
                improved.  1 indicates an improvement.
        TUCE  - Test score on economics test
        PSI   - participation in program
        GPA   - Student's grade point average

    GPA  TUCE  PSI
0  2.66  20.0  0.0
1  2.89  22.0  0.0
2  3.28  24.0  0.0
3  2.92  12.0  0.0
4  4.00  21.0  0.0

Them, we add a constant to the matrix of regressors:

In [3]:
exog = sm.add_constant(exog, prepend=True)

To create your own Likelihood Model, you simply need to overwrite the loglike method.

In [4]:
class MyProbit(GenericLikelihoodModel):
    def loglike(self, params):
        exog = self.exog
        endog = self.endog
        q = 2 * endog - 1
        return stats.norm.logcdf(q*np.dot(exog, params)).sum()

Estimate the model and print a summary:

In [5]:
sm_probit_manual = MyProbit(endog, exog).fit()
print(sm_probit_manual.summary())
Optimization terminated successfully.
         Current function value: 0.400588
         Iterations: 292
         Function evaluations: 494
                               MyProbit Results                               
==============================================================================
Dep. Variable:                  GRADE   Log-Likelihood:                -12.819
Model:                       MyProbit   AIC:                             33.64
Method:            Maximum Likelihood   BIC:                             39.50
Date:                Mon, 14 May 2018                                         
Time:                        21:44:55                                         
No. Observations:                  32                                         
Df Residuals:                      28                                         
Df Model:                           3                                         
==============================================================================
                 coef    std err          z      P>|z|      [0.025      0.975]
------------------------------------------------------------------------------
const         -7.4523      2.542     -2.931      0.003     -12.435      -2.469
GPA            1.6258      0.694      2.343      0.019       0.266       2.986
TUCE           0.0517      0.084      0.617      0.537      -0.113       0.216
PSI            1.4263      0.595      2.397      0.017       0.260       2.593
==============================================================================

Compare your Probit implementation to statsmodels' "canned" implementation:

In [6]:
sm_probit_canned = sm.Probit(endog, exog).fit()
Optimization terminated successfully.
         Current function value: 0.400588
         Iterations 6
In [7]:
print(sm_probit_canned.params)
print(sm_probit_manual.params)
const   -7.452320
GPA      1.625810
TUCE     0.051729
PSI      1.426332
dtype: float64
[-7.45233176  1.62580888  0.05172971  1.42631954]
In [8]:
print(sm_probit_canned.cov_params())
print(sm_probit_manual.cov_params())
          const       GPA      TUCE       PSI
const  6.464166 -1.169668 -0.101173 -0.594792
GPA   -1.169668  0.481473 -0.018914  0.105439
TUCE  -0.101173 -0.018914  0.007038  0.002472
PSI   -0.594792  0.105439  0.002472  0.354070
[[ 6.46416770e+00 -1.16966617e+00 -1.01173181e-01 -5.94789009e-01]
 [-1.16966617e+00  4.81472117e-01 -1.89134591e-02  1.05438228e-01]
 [-1.01173181e-01 -1.89134591e-02  7.03758403e-03  2.47189233e-03]
 [-5.94789009e-01  1.05438228e-01  2.47189233e-03  3.54069514e-01]]

Notice that the GenericMaximumLikelihood class provides automatic differentiation, so we didn't have to provide Hessian or Score functions in order to calculate the covariance estimates.

Example 2: Negative Binomial Regression for Count Data

Consider a negative binomial regression model for count data with log-likelihood (type NB-2) function expressed as:

$$ \mathcal{L}(\beta_j; y, \alpha) = \sum_{i=1}^n y_i ln \left ( \frac{\alpha exp(X_i'\beta)}{1+\alpha exp(X_i'\beta)} \right ) - \frac{1}{\alpha} ln(1+\alpha exp(X_i'\beta)) + ln \Gamma (y_i + 1/\alpha) - ln \Gamma (y_i+1) - ln \Gamma (1/\alpha) $$

with a matrix of regressors $X$, a vector of coefficients $\beta$, and the negative binomial heterogeneity parameter $\alpha$.

Using the nbinom distribution from scipy, we can write this likelihood simply as:

In [9]:
import numpy as np
from scipy.stats import nbinom
In [10]:
def _ll_nb2(y, X, beta, alph):
    mu = np.exp(np.dot(X, beta))
    size = 1/alph
    prob = size/(size+mu)
    ll = nbinom.logpmf(y, size, prob)
    return ll

New Model Class

We create a new model class which inherits from GenericLikelihoodModel:

In [11]:
from statsmodels.base.model import GenericLikelihoodModel
In [12]:
class NBin(GenericLikelihoodModel):
    def __init__(self, endog, exog, **kwds):
        super(NBin, self).__init__(endog, exog, **kwds)
        
    def nloglikeobs(self, params):
        alph = params[-1]
        beta = params[:-1]
        ll = _ll_nb2(self.endog, self.exog, beta, alph)
        return -ll 
    
    def fit(self, start_params=None, maxiter=10000, maxfun=5000, **kwds):
        # we have one additional parameter and we need to add it for summary
        self.exog_names.append('alpha')
        if start_params == None:
            # Reasonable starting values
            start_params = np.append(np.zeros(self.exog.shape[1]), .5)
            # intercept
            start_params[-2] = np.log(self.endog.mean())
        return super(NBin, self).fit(start_params=start_params, 
                                     maxiter=maxiter, maxfun=maxfun, 
                                     **kwds) 

Two important things to notice:

  • nloglikeobs: This function should return one evaluation of the negative log-likelihood function per observation in your dataset (i.e. rows of the endog/X matrix).
  • start_params: A one-dimensional array of starting values needs to be provided. The size of this array determines the number of parameters that will be used in optimization.

That's it! You're done!

Usage Example

The Medpar dataset is hosted in CSV format at the Rdatasets repository. We use the read_csv function from the Pandas library to load the data in memory. We then print the first few columns:

In [13]:
import statsmodels.api as sm
In [14]:
medpar = sm.datasets.get_rdataset("medpar", "COUNT", cache=True).data

medpar.head()
Out[14]:
los hmo white died age80 type type1 type2 type3 provnum
0 4 0 1 0 0 1 1 0 0 30001
1 9 1 1 0 0 1 1 0 0 30001
2 3 1 1 1 1 1 1 0 0 30001
3 9 0 1 0 0 1 1 0 0 30001
4 1 0 1 1 1 1 1 0 0 30001

The model we are interested in has a vector of non-negative integers as dependent variable (los), and 5 regressors: Intercept, type2, type3, hmo, white.

For estimation, we need to create two variables to hold our regressors and the outcome variable. These can be ndarrays or pandas objects.

In [15]:
y = medpar.los
X = medpar[["type2", "type3", "hmo", "white"]].copy()
X["constant"] = 1

Then, we fit the model and extract some information:

In [16]:
mod = NBin(y, X)
res = mod.fit()
Optimization terminated successfully.
         Current function value: 3.209014
         Iterations: 805
         Function evaluations: 1238

Extract parameter estimates, standard errors, p-values, AIC, etc.:

In [17]:
print('Parameters: ', res.params)
print('Standard errors: ', res.bse)
print('P-values: ', res.pvalues)
print('AIC: ', res.aic)
Parameters:  [ 0.2212642   0.70613942 -0.06798155 -0.12903932  2.31026565  0.44575147]
Standard errors:  [0.05059259 0.07613047 0.05326097 0.06854141 0.06794692 0.01981542]
P-values:  [1.22297920e-005 1.76978024e-020 2.01819161e-001 5.97481600e-002
 2.15079626e-253 4.62688439e-112]
AIC:  9604.95320583016

As usual, you can obtain a full list of available information by typing dir(res). We can also look at the summary of the estimation results.

In [18]:
print(res.summary())
                                 NBin Results                                 
==============================================================================
Dep. Variable:                    los   Log-Likelihood:                -4797.5
Model:                           NBin   AIC:                             9605.
Method:            Maximum Likelihood   BIC:                             9632.
Date:                Mon, 14 May 2018                                         
Time:                        21:44:58                                         
No. Observations:                1495                                         
Df Residuals:                    1490                                         
Df Model:                           4                                         
==============================================================================
                 coef    std err          z      P>|z|      [0.025      0.975]
------------------------------------------------------------------------------
type2          0.2213      0.051      4.373      0.000       0.122       0.320
type3          0.7061      0.076      9.275      0.000       0.557       0.855
hmo           -0.0680      0.053     -1.276      0.202      -0.172       0.036
white         -0.1290      0.069     -1.883      0.060      -0.263       0.005
constant       2.3103      0.068     34.001      0.000       2.177       2.443
alpha          0.4458      0.020     22.495      0.000       0.407       0.485
==============================================================================

Testing

We can check the results by using the statsmodels implementation of the Negative Binomial model, which uses the analytic score function and Hessian.

In [19]:
res_nbin = sm.NegativeBinomial(y, X).fit(disp=0)
print(res_nbin.summary())
                     NegativeBinomial Regression Results                      
==============================================================================
Dep. Variable:                    los   No. Observations:                 1495
Model:               NegativeBinomial   Df Residuals:                     1490
Method:                           MLE   Df Model:                            4
Date:                Mon, 14 May 2018   Pseudo R-squ.:                 0.01215
Time:                        21:44:58   Log-Likelihood:                -4797.5
converged:                       True   LL-Null:                       -4856.5
                                        LLR p-value:                 1.404e-24
==============================================================================
                 coef    std err          z      P>|z|      [0.025      0.975]
------------------------------------------------------------------------------
type2          0.2212      0.051      4.373      0.000       0.122       0.320
type3          0.7062      0.076      9.276      0.000       0.557       0.855
hmo           -0.0680      0.053     -1.276      0.202      -0.172       0.036
white         -0.1291      0.069     -1.883      0.060      -0.263       0.005
constant       2.3103      0.068     34.001      0.000       2.177       2.443
alpha          0.4457      0.020     22.495      0.000       0.407       0.485
==============================================================================
In [20]:
print(res_nbin.params)
type2       0.221218
type3       0.706173
hmo        -0.067987
white      -0.129053
constant    2.310279
alpha       0.445748
dtype: float64
In [21]:
print(res_nbin.bse)
type2       0.050592
type3       0.076131
hmo         0.053261
white       0.068541
constant    0.067947
alpha       0.019815
dtype: float64

Or we could compare them to results obtained using the MASS implementation for R:

url = 'http://vincentarelbundock.github.com/Rdatasets/csv/COUNT/medpar.csv'
medpar = read.csv(url)
f = los~factor(type)+hmo+white

library(MASS)
mod = glm.nb(f, medpar)
coef(summary(mod))
                 Estimate Std. Error   z value      Pr(>|z|)
(Intercept)    2.31027893 0.06744676 34.253370 3.885556e-257
factor(type)2  0.22124898 0.05045746  4.384861  1.160597e-05
factor(type)3  0.70615882 0.07599849  9.291748  1.517751e-20
hmo           -0.06795522 0.05321375 -1.277024  2.015939e-01
white         -0.12906544 0.06836272 -1.887951  5.903257e-02

Numerical precision

The statsmodels generic MLE and R parameter estimates agree up to the fourth decimal. The standard errors, however, agree only up to the second decimal. This discrepancy is the result of imprecision in our Hessian numerical estimates. In the current context, the difference between MASS and statsmodels standard error estimates is substantively irrelevant, but it highlights the fact that users who need very precise estimates may not always want to rely on default settings when using numerical derivatives. In such cases, it is better to use analytical derivatives with the LikelihoodModel class.

© 2009–2012 Statsmodels Developers
© 2006–2008 Scipy Developers
© 2006 Jonathan E. Taylor
Licensed under the 3-clause BSD License.
http://www.statsmodels.org/stable/examples/notebooks/generated/generic_mle.html