CODEWITHSUNDEEP
pythonpandas
xena
xena

Reputation: 15

How do I display correlation coefficients of each individual variable in Python?

I ran a linear regression model and have my coefficients. How do I print my variables next to my coefficients?

df = pd.read_csv('data', sep=";")
reg = linear_model.LinearRegression()
reg.fit(df[["age", "area", "bedrooms"]],df.price)
print(reg.coef_)

Output 

[ 6.55199614e-02 -1.86317709e+00  2.20902007e-02]

I want the output to be

age coef: 6.55199614e-02
area coef: -1.86317709e+00
bedroom coef: 2.20902007e-02

Upvotes: 0

Views: 785

Answers (3)

imak
imak

Reputation: 411

if you just want to print your variables next to your coefficients, you can use something like that:

df = pd.read_csv('data', sep=";")
reg = linear_model.LinearRegression()

colnames = ["age", "area", "bedrooms"]

reg.fit(df[colnames],df.price)

coefs_map = dict(zip(colnames, reg.coef_))

for k,v in coefs_map.items():
    print(f"{k}: {v}")

Upvotes: 0

jedi
jedi

Reputation: 585

My preferred approach is pd.Series(reg.coef_,index=df.columns), then printing comes for free. Also it is easier to work with pd.Series for other calculations, comparisons of models via pd.concat etc.

Upvotes: 1

Acccumulation
Acccumulation

Reputation: 3591

I believe that '\n'.join(col+' coef: '+str(coef) for (col, coef) in zip(df.columns, reg.coef_)) gets you what you asked for. However, I recommend the following, as it gives output that is easier to read:

pad_length = 8+max(len(col) for col in df.columns)
output = '\n'.join((col+' coef: ').ljust(pad_length)+str(coef) for 
    (col, coef) in zip(df.columns, reg.coef_))

Also, all of this assumes that your column names are all already strings. If they're not, you should replace col with str(col).

Upvotes: 0

Related Questions