Plot sklearn LinearRegression output with matplotlib

Question

After importing the file when I separate the x_values and y_values using numpy as:

import pandas as pd
from sklearn import linear_model
from  matplotlib import pyplot 
import numpy as np

#read data
dataframe = pd.read_csv('challenge_dataset.txt')
dataframe.columns=['Brain','Body']
x_values=np.array(dataframe['Brain'],dtype=np.float64).reshape(1,-1)
y_values=np.array(dataframe['Body'],dtype=np.float64).reshape(1,-1)

#train model on data
body_reg = linear_model.LinearRegression()
body_reg.fit(x_values, y_values)
prediction=body_reg.predict(x_values)

print(prediction)
#visualize results
pyplot.scatter(x_values, y_values)
pyplot.plot(x_values,prediction)
pyplot.show()

I get the plot as following image, which doesn't show up the line of best fit and also when I print the value of 'prediction' it shows up values same as 'y_values'.

Contrary when I use the following code. I get the regression line.

#read data
dataframe = pd.read_csv('challenge_dataset.txt')
dataframe.columns=['Brain','Body']
x_values=dataframe[['Brain']]
y_values=dataframe[['Body']]

Why is it so ?

Thanks in advance.

x_values=np.array(dataframe['Brain'],dtype=np.float64).reshape(1,-1) Because I was taking the value of column Brain in 1 dimension. I know it's weird I could have taken it in 2 dimension but I was just experimenting. — Mayank Raj
– Mayank Raj, Commented Sep 23, 2017 at 18:08
What I mean is what happens if you leave .reshape(1,-1) out? — ImportanceOfBeingErnest
– ImportanceOfBeingErnest, Commented Sep 23, 2017 at 18:12
It throws this error. ValueError: Expected 2D array, got 1D array instead: — Mayank Raj
– Mayank Raj, Commented Sep 23, 2017 at 18:14

ImportanceOfBeingErnest · Accepted Answer · 2017-09-23 18:39:07Z

linear_model.LinearRegression().fit(X,y) expects its arguments

X : numpy array or sparse matrix of shape [n_samples,n_features]
y : numpy array of shape [n_samples, n_targets]

Here you have 1 "feature" and 1 "target", hence the expected shape of the input would be (n_samples,1)

While this is the case for

x_values=dataframe[['Brain']]
y_values=dataframe[['Body']]

the shape for np.array(dataframe['Brain'],dtype=np.float64).reshape(1,-1) is (n_samples,).

Another option to optain the desired shape from the dataframe columns would be to broadcast them to a 2D array with a new axis

x_values=dataframe['Brain'].values[:,np.newaxis]
y_values=dataframe['Body'].values[:,np.newaxis]

Note that in order to show a nice line, you would probably want to sort the x values.

import pandas as pd
from sklearn import linear_model
from  matplotlib import pyplot 
import numpy as np

#read data
x = np.random.rand(25,2)
x[:,1] = 2*x[:,0]+np.random.rand(25)
dataframe = pd.DataFrame(x,columns=['Brain','Body'])


x_values=dataframe['Brain'].values[:,np.newaxis]
y_values=dataframe['Body'].values[:,np.newaxis]

body_reg = linear_model.LinearRegression()
body_reg.fit(x_values, y_values)
prediction=body_reg.predict(np.sort(x_values, axis=0))

pyplot.scatter(x_values, y_values)
pyplot.plot(np.sort(x_values, axis=0),prediction)
pyplot.show()

Collectives™ on Stack Overflow

Plot sklearn LinearRegression output with matplotlib

1 Answer 1

Comments

Your Answer

Linked

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

Comments

Your Answer

Sign up or log in

Post as a guest

Linked

Related