0
from sklearn import linear_model
regr = linear_model.LinearRegression()
x = np.asanyarray(train[['ENGINESIZE','CYLINDERS','FUELCONSUMPTION_COMB']])
y = np.asanyarray(train[['CO2EMISSIONS']])
regr.fit (x, y)
# The coefficients
print ('Coefficients: ', regr.coef_)
x1 = np.asanyarray(test[['ENGINESIZE','CYLINDERS','FUELCONSUMPTION_COMB']])
y1 = np.asanyarray(test[['CO2EMISSIONS']])
xy = regr.predict(y1)
print(xy) // an error is generating while printing this (valueError)

this worked in simple linear regression but here is not working in multiple-linear-regression

likith vishal
  • 1

2 Answers2

1

regr.predict expects the same shape of x.

Furthermore, when you want to predict something, it should be based on some input, not output.

So, xy = regr.predict(y1) is wrong.

You should try xy = regr.predict(x1) instead.

Pedro Miguel
  • 66
  • 3
0

The reason why it works (but in fact, it is not correct) in simple regression in that you provide a 1D array to regr.predict. As mentionned, this should be regr.predict(x1) instread of regr.predict(y1), since you are trying to predict y1 from x1. The algorithm does not "distinguish" between x1 and y1 in simple regression because they are both 1D arrays, so it does not raise an error.

However in multiple regression, you fit an equation on a 2D or 3D or...N-dimensional x array. So, when you run regr.predict(y1), it raises an error because you are trying to predict with the 1D y1 array.

Just replace regr.predict(y1) by regr.predict(x1) and it will work both for simple and multiple regrerssion.

G.L
  • 128
  • 8