Bdtjtk

I'm new to machine learning and I'm trying out simple linear regression for the first time on my pc. I am getting an error saying value error expected 2d array but got 1d array instead.

I have no clue on what to do. Any tips in fixing this code will be helpful.

import numpy as np

import matplotlib as plt

import pandas as pd

#preparing dataset

dataset = pd.read_csv("Salary_Data.csv")

X = dataset.iloc[:,0].values

Y = dataset.iloc[:,1].values

#Missing values

#Encoding

#splitting Dataset

from sklearn.model_selection import train_test_split

X_train,X_test,Y_train,Y_test = train_test_split(X,Y, test_size = 1/3, random_state = 0)

#Format Scaling



#Simple Linear regressing

from sklearn.linear_model import LinearRegression

regressor = LinearRegression()

regressor.fit(X_train,Y_train)

The error I got is:

regressor.fit(X_train,Y_train)

Traceback (most recent call last):



  File "<ipython-input-10-4d17c24ccad2>", line 1, in <module>

    regressor.fit(X_train,Y_train)



  File "C:UsershomeAnacondalibsite-packagessklearnlinear_modelbase.py", line 458, in fit

    y_numeric=True, multi_output=True)



  File "C:UsershomeAnacondalibsite-packagessklearnutilsvalidation.py", line 756, in check_X_y

    estimator=estimator)



  File "C:UsershomeAnacondalibsite-packagessklearnutilsvalidation.py", line 552, in check_array

    "if it contains a single sample.".format(array))



ValueError: Expected 2D array, got 1D array instead:

array=[ 2.9  5.1  3.2  4.5  8.2  6.8  1.3 10.5  3.   2.2  5.9  6.   3.7  3.2

  9.   2.   1.1  7.1  4.9  4. ].

Reshape your data either using array.reshape(-1, 1) if your data has a single feature or array.reshape(1, -1) if it contains a single sample.

Per the sklearn docs LinearRegression::fit() expects that the shape of X is (n_samples, n_features) and the shape of Y is (n_samples, n_targets). In the case in question I believe this would correspond to (20L, 1) and (20L, ) (i.e. one 'row vector' of length 20 and one 'column' vector of length 20). The call

X = dataset.iloc[:,0].values

Produces an object shaped (20L, ). So this needs to be reshaped to (20L, 1) which is fairly trivial (and actually what the somewhat cryptic suggestion "use array.reshape(-1, 1)" is trying to tell you),

X = dataset.iloc[:,0].values.reshape(-1, 1)

as @Dan has noted, an appropriate alternative is

X = dataset.iloc[:,[0]].values

Both options will produce an object of shape (20L, 1) which is then compatible with the Y object of shape (20L, ). Note that the shapes of X_train and Y_train shadow those of X and Y, respectively.