Error: Found array with dim 3. Estimator expected <= 2

Question

I have a 14x5 data matrix titled data. The first column (Y) is the dependent variable followed by 4 independent variables (X,S1,S2,S3). When trying to fit a regression model to a subset of the independent variables ['S2'][:T] I get the following error:

ValueError: Found array with dim 3. Estimator expected <= 2.

I'd appreciate any insight on a fix. Code below.

import pandas as pd
import numpy as np
from sklearn.linear_model import LinearRegression


data = pd.read_csv('C:/path/Macro.csv')
T=len(data['X'])-1

#Fit variables
X = data['X'][:T]
S1 = data['S1'][:T]
S2 = data['S2'][:T]
S3 = data['S3'][:T]
Y = data['Y'][:T]

regressor = LinearRegression()
regressor.fit([[X,S1,S2,S3]], Y)

Is [this question](http://stackoverflow.com/questions/34436828/how-to-solve-several-independent-time-series-at-the-same-time-using-scikit-linea) any help? — Prune, Jan 19 '16 at 01:15

Igor Raush · Accepted Answer · 2016-01-19T02:17:25.007

You are passing a 3-dimensional array as the first argument to fit(). X, S1, S2, S3 are all Series objects (1-dimensional), so the following

[[X, S1, S2, S3]]

is 3-dimensional. sklearn estimators expect an array of feature vectors (2-dimensional).

Try something like this:

# pandas indexing syntax
# data.ix[ row index/slice, column index/slice ]

X = data.ix[:T, 'X':]  # rows up to T, columns from X onward
y = data.ix[:T, 'Y']   # rows up to T, Y column
regressor = LinearRegression()
regressor.fit(X, y)

Error: Found array with dim 3. Estimator expected <= 2

1 Answers1

Linked