3
$\begingroup$

I'm trying to get to grips with sci-kit learn for some simple machine learning projects but I'm coming unstuck with Pipelines and wonder what I've done wrong...

I'm trying to work through a tutorial on Kaggle

Here's my code:

import pandas as pd

train = pd.read_csv(local path to training data) train_labels =
pd.read_csv(local path to labels)


from sklearn.decomposition import PCA
from sklearn.svm import LinearSVC
from sklearn.grid_search import GridSearchCV

pca = PCA()
clf = LinearSVC()

n_components = arange(1, 39)
loss = ['l1','l2']
penalty = ['l1','l2']
C = arange(0, 1, .1)
whiten = [True, False]

from sklearn.pipeline import Pipeline

#set up pipeline
pipe = Pipeline(steps=[('pca', pca), ('clf', clf)])

#set up GridsearchCV
estimator = GridSearchCV(pipe, dict(pca__n_components = n_components, pca__whiten = whiten, clf__loss = loss, clf__penalty = penalty, clf__C = C)) 

> estimator

Returns:

GridSearchCV(cv=None,
       estimator=Pipeline(steps=[('pca', PCA(copy=True, n_components=None, whiten=False)), ('clf', LinearSVC(C=1.0, class_weight=None, dual=True, fit_intercept=True,
     intercept_scaling=1, loss='l2', multi_class='ovr', penalty='l2',
     random_state=None, tol=0.0001, verbose=0))]),
       fit_params={}, iid=True, loss_func=None, n_jobs=1,
       param_grid={'clf__penalty': ['l1', 'l2'], 'clf__loss': ['l1', 'l2'], 'clf__C': array([ 0. ,  0.1,  0.2,  0.3,  0.4,  0.5,  0.6,  0.7,  0.8,  0.9]), 'pca__n_components': array([ 1,  2,  3,  4,  5,  6,  7,  8,  9, 10, 11, 12, 13, 14, 15, 16, 17,
       18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34,
       35, 36, 37, 38]), 'pca__whiten': [True, False]},
       pre_dispatch='2*n_jobs', refit=True, score_func=None, scoring=None,
       verbose=0)

But when I try to train data:

estimator.fit(train, train_labels)
The error is:

    428         for test_fold_idx, per_label_splits in enumerate(zip(*per_label_cvs)):
    429             for label, (_, test_split) in zip(unique_labels, per_label_splits):
--> 430                 label_test_folds = test_folds[y == label]
    431                 # the test split can be too big because we used
    432                 # KFold(max(c, self.n_folds), self.n_folds) instead of

IndexError: too many indices for array

Can anyone point me in the right direction?

$\endgroup$
1
$\begingroup$

It turns out that the Pandas dataframe is the wrong shape.

estimator.fit(train.values, train_labels[0].values)

works, although I also had to drop the penalty term.

| improve this answer | |
$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.