Fitting sklearn's SVM classifier with data from a pandas DataFrame

user_7021584 0
Viewed 9462

I'm trying to use SVM but I dont know how to fit the model when I am using pandas data frame. If my data looks like this:

df = pd.DataFrame({"x": ['011', '100', '111'] , "y": [0,1,0]})
df.x.apply(lambda x: np.array(list(map(int,x))))
>>>df
    x           y
0   [0, 1, 1]   0
1   [1, 0, 0]   1
2   [1, 1, 1]   0

If I try to fit the model this way:

clf = svm.SVC().fit(df.x, df.y)

I am getting this error:

ValueError: setting an array element with a sequence.

What is the correct way to fit the SVM using this data frame?

scikit-learn pandas dataframe numpy python
3 Answers
df = pd.DataFrame({"x": ['011', '100', '111'] , "y": [0,1,0]})
df.x = df.x.apply(lambda x: list(map(int,x)))

df
           x  y
0  [0, 1, 1]  0
1  [1, 0, 0]  1
2  [1, 1, 1]  0

df.x is a column of arrays. This probably isn't the best way to store data, and it would seem sklearn isn't very good at understanding it. It would be simpler to convert everything to a list of lists and pass that to SVC. Try this:

x = df.x.tolist()
print(x)
[[0, 1, 1], [1, 0, 0], [1, 1, 1]]
SVC().fit(x, df.y)
user_49090870

Another solution is the code below.

import pandas as pd
import numpy as np

from sklearn.svm import SVC

df = pd.DataFrame({"x": ['011', '100', '111'] , "y": [0,1,0]})
x = df.x.apply(lambda x: pd.Series(list(x)))
x
# Out[2]:
#    0  1  2
# 0  0  1  1
# 1  1  0  0
# 2  1  1  1

SVC().fit(x, df.y)
# Out[3]:
# SVC(C=1.0, cache_size=200, class_weight=None, coef0=0.0,
#   decision_function_shape=None, degree=3, gamma='auto', kernel='rbf',
#   max_iter=-1, probability=False, random_state=None, shrinking=True,
#   tol=0.001, verbose=False)
user_45104210 
import numpy as np
from sklearn.svm import SVC

# Convert your data frame's columns into arrays
features = df['x'].to_numpy()
labels = df['y'].to_numpy()

# feed into your classifier 
SVC().fit(features,labels)
user_148287000
Related