Hoe pas je een GLM in Python?

Generalized Linear Models in Python

Ita Cirovic Donev

Data Science Consultant

statsmodels

statsmodels importeren
```
import statsmodels.api as sm
```
Ondersteuning voor formules
```
import statsmodels.formula.api as smf
```

glm() direct gebruiken

from statsmodels.formula.api import glm

Proces van model fitten

Beschrijf het model $\rightarrow \texttt{\color{#007AFF}{glm()}}$
Fit het model $\rightarrow \texttt{\color{#007AFF}{.fit()}}$
Vat het model samen $\rightarrow \texttt{\color{#007AFF}{.summary()}}$
Maak voorspellingen $\rightarrow \texttt{\color{#007AFF}{.predict()}}$

Het model beschrijven

OP FORMULE gebaseerd

from statsmodels.formula.api import glm

model = glm(formule, data, family)

OP ARRAY gebaseerd

import statsmodels.api as sm

X = sm.add_constant(X)
model = sm.glm(y, X, family)

Formula-argument

$$\texttt{\color{#00A388}{respons}} \sim \texttt{\color{#FF6138}{verklarende variabele(n)}}$$ $$\texttt{\color{#00A388}{output}} \sim \texttt{\color{#FF6138}{input(s)}}$$

formula = 'y ~ x1 + x2'

$\texttt{\color{#FF6138}{C(x1)}}$ : behandel x1 als categorische variabele
$\texttt{\color{#FF6138}{-1}}$ : verwijder intercept
$\texttt{\color{#FF6138}{x1:x2}}$ : interactieterm tussen x1 en x2
$\texttt{\color{#FF6138}{x1*x2}}$ : interactieterm en de losse variabelen
$\texttt{\color{#FF6138}{np.log(x1)}}$ : pas vectorfuncties toe op variabelen

Family-argument

family = sm.families.____()

De family-functies:

$\texttt{\color{#007AFF}{Gaussian}(link = sm.families.links.\color{deeppink}{identity()})}$ $\rightarrow$ de standaard family
$\texttt{\color{#007AFF}{Binomial}(link = sm.families.links.\color{deeppink}{logit()})}$
- $\texttt{\color{deeppink}{probit()}}$, $\texttt{\color{deeppink}{cauchy()}}$, $\texttt{\color{deeppink}{log()}}$, en $\texttt{\color{deeppink}{cloglog()}}$
$\texttt{\color{#007AFF}{Poisson}(link = sm.families.links.\color{deeppink}{log()})}$
- $\texttt{\color{deeppink}{identity()}}$ en $\texttt{\color{deeppink}{sqrt()}}$

Andere distributiefamilies vind je op de statsmodels-website.

Model samenvatten

print(model_GLM.summary())

                 Generalized Linear Model Regression Results                  
=============================================================================
Dep. Variable:                      y  No. Observations:                  173
Model:                            GLM  Df Residuals:                      171
Model Family:                Binomial  Df Model:                            1
Link Function:                  logit  Scale:                          1.0000
Method:                          IRLS  Log-Likelihood:                -97.226
Date:                Mon, 21 Jan 2019  Deviance:                       194.45
Time:                        11:30:01  Pearson chi2:                     165.
No. Iterations:                     4  Covariance Type:             nonrobust
=============================================================================
                 coef    std err         z      P>|z|      [0.025      0.975]
-----------------------------------------------------------------------------
Intercept    -12.3508      2.629    -4.698      0.000     -17.503      -7.199
width          0.4972      0.102     4.887      0.000       0.298       0.697
=============================================================================

Regressiecoëfficiënten

$\texttt{\color{#007AFF}{.params}}$ print regressiecoëfficiënten

model_GLM.params

Intercept   -12.350818
width         0.497231
dtype: float64

$\texttt{\color{#007AFF}{.conf\_int(alpha=0.05, cols=None)}}$ print betrouwbaarheidsintervallen

model_GLM.conf_int()

                   0         1
Intercept -17.503010 -7.198625
width       0.297833  0.696629

Voorspellingen

Specificeer alle modelvariabelen in testdata
$\texttt{\color{#007AFF}{.predict(\color{#FF931B}{test\_data})}}$ berekent voorspellingen

model_GLM.predict(test_data)

0    0.029309
1    0.470299
2    0.834983
3    0.972363
4    0.987941

Laten we oefenen!

Generalized Linear Models in Python