CODEWITHSUNDEEP
pythonpandasdataframescikit-learnlinear-regression
Luigi87
Luigi87

Reputation: 275

how to create a dataframe with beta coefficients of regression

I have the following code, which takes each column of the data frame HF and fit it to the entire dataframe Index. What I am interested then is to produce a dataframe Betas which contains the regression coefficients:

col_names = HF.columns
Betas = pd.DataFrame(columns= HF.columns, index = Index.columns)
for name in col_names:
    lm = linear_model.LinearRegression()
    hf = HF[name]
    model = lm.fit(Index, hf)
    Betas.loc[:][name] = lm.coef_

the code runs but what I get is a Betas dataframe with only NaN values. Can you please help understand the problem? Thanks

Upvotes: 0

Views: 1086

Answers (1)

Ric S
Ric S

Reputation: 9257

The problem is that after .loc you are slicing the Betas dataframe twice, whereas you need to do it only once by using the same set of squared brackets.

This should be your new for-cycle

for name in col_names:
    lm = linear_model.LinearRegression()
    hf = HF[name]
    lm.fit(Index, hf)
    Betas.loc[:, name] = lm.coef_

As a sidenote, I removed the assignment to model as it is not necessary; you can just call the fit method on the lm object.

Upvotes: 1

Related Questions