Value Error using Patsy for a Logistic Regression Model

Question

I have made an extremely simple logistic model for the purpose of my question. Here is the code below;

import patsy
import pandas as pd
import statsmodels.api as sm

df = pd.DataFrame()

for i in range(5):
    df.at[i, 'response'] = 1
    if i == 3:
        df.at[i,'response'] = 0

df['x'] = range(5)

y, X = patsy.dmatrices('response ~ x', df,return_type = 'dataframe')
logit_model=sm.Logit(y,X)
result=logit_model.fit()

ypred = logit_model.predict(X)
print(ypred)

Please excuse my crappy code, i'm writing this in a rush - need to go to work haha. This code is throwing up a value error - ValueError: shapes (5,2) and (5,2) not aligned: 2 (dim 1) != 5 (dim 0) about line 18.

I genuinely don't understand how these are not aligned as I am simply passing through the train data X back into the model using predict(). My feeling is that I am missing something about patsy.dmatrices.

Anyone have an idea?

Answer 1

You assigned the fitted object to result, so you should use that to predict:

result.predict(X)

To get the fitted values, you can also do:

result.fittedvalues