Exclude values from polynomial fit in pandas-CodePudding

I would like to exclude certain values from a polynomial fit to data in a pandas data frame. At the moment my code is:

d = np.polyfit(df1['S_B'],df1['log_norm'],1)
f = np.poly1d(d)
df1['predicted']=getExp(f(df1['S_B']))*df1['Cl']

Is it possible to exclude the values in column log_norm at rows 0,1 and 6 in the polynomial fit?

The code to get the getExp function is below

from math import e
def getExp(n):
 return e**n

CodePudding user response：

I would just create a new dataframe where you have dropped the rows you don't want:

df2 = df1.drop(index=[0, 1, 6])

An alternative would be to make these rows NaN in the log_norm column, but np.polyfit might throw an error, in which case you would need to drop these rows anyway.