Welcome to expandas’s documentation!¶

Contents:

Data Handling¶

Data Preparation¶

This section describes how to prepare basic data format named ModelFrame. ModelFrame defines a metadata to specify target (response variable) and data (explanatory variable / features). Using these metadata, ModelFrame can call other statistics/ML functions in more simple way.

You can create ModelFrame as the same manner as pandas.DataFrame. The below example shows how to create basic ModelFrame, which DOESN’T have target values.

>>> import expandas as expd

>>> df = expd.ModelFrame({'A': [1, 2, 3], 'B': [2, 3, 4],
...                       'C': [3, 4, 5]}, index=['A', 'B', 'C'])
>>> df
   A  B  C
A  1  2  3
B  2  3  4
C  3  4  5

>>> type(df)
<class 'expandas.core.frame.ModelFrame'>

You can check whether the created ModelFrame has target values using ModelFrame.has_target() function.

>>> df.has_target()
False

Target values can be specifyied via target keyword. You can simply pass a column name to be handled as target. Target column name can be confirmed via target_name property.

>>> df2 = expd.ModelFrame({'A': [1, 2, 3], 'B': [2, 3, 4],
...                        'C': [3, 4, 5]}, target='A')
>>> df2
   A  B  C
0  1  2  3
1  2  3  4
2  3  4  5

>>> df2.has_target()
True

>>> df2.target_name
'A'

Also, you can pass any list-likes to be handled as a target. In this case, target column will be named as ”.target”.

>>> df3 = expd.ModelFrame({'A': [1, 2, 3], 'B': [2, 3, 4],
...                        'C': [3, 4, 5]}, target=[4, 5, 6])
>>> df3
   .target  A  B  C
0        4  1  2  3
1        5  2  3  4
2        6  3  4  5

>>> df3.has_target()
True

>>> df3.target_name
'.target'

Also, you can pass pandas.DataFrame and pandas.Series as data and target.

>>> import pandas as pd
df4 = expd.ModelFrame({'A': [1, 2, 3], 'B': [2, 3, 4],
...                    'C': [3, 4, 5]}, target=pd.Series([4, 5, 6]))
>>> df4
   .target  A  B  C
0        4  1  2  3
1        5  2  3  4
2        6  3  4  5

>>> df4.has_target()
True

>>> df4.target_name
'.target'

Note

Target values are mandatory to perform operations which require response variable, such as regression and supervised learning.

Data Manipulation¶

You can access to each property as the same as pandas.DataFrame. Sliced results will be ModelSeries (simple wrapper for pandas.Series to support some data manipulation) or ModelFrame

>>> df
   A  B  C
A  1  2  3
B  2  3  4
C  3  4  5

>>> sliced = df['A']
>>> sliced
A    1
B    2
C    3
Name: A, dtype: int64

>>> type(sliced)
<class 'expandas.core.series.ModelSeries'>

>>> subset = df[['A', 'B']]
>>> subset
   A  B
A  1  2
B  2  3
C  3  4

>>> type(subset)
<class 'expandas.core.frame.ModelFrame'>

ModelFrame has a special properties data to access data (features) and target to access target.

>>> df2
   A  B  C
0  1  2  3
1  2  3  4
2  3  4  5

>>> df2.target_name
'A'

>>> df2.data
   B  C
0  2  3
1  3  4
2  4  5

>>> df2.target
0    1
1    2
2    3
Name: A, dtype: int64

You can update data and target via properties, in addition to standard pandas.DataFrame ways.

>>> df2.target = [9, 9, 9]
>>> df2
   A  B  C
0  9  2  3
1  9  3  4
2  9  4  5

>>> df2.data = pd.DataFrame({'X': [1, 2, 3], 'Y': [4, 5, 6]})
>>> df2
   A  X  Y
0  9  1  4
1  9  2  5
2  9  3  6

>>> df2['X'] = [0, 0, 0]
>>> df2
   A  X  Y
0  9  0  4
1  9  0  5
2  9  0  6

You can change target column specifying target_name property. Specifying a column which doesn’t exist in ModelFrame results in target column to be data column.

>>> df2.target_name
'A'

>>> df2.target_name = 'X'
>>> df2.target_name
'X'

>>> df2.target_name = 'XXXX'
>>> df2.has_target()
False

>>> df2.data
   A  X  Y
0  9  0  4
1  9  0  5
2  9  0  6

Use scikit-learn¶

This section describes how to use scikit-learn functionalities via expandas.

Basics¶

You can create ModelFrame instance from scikit-learn datasets directly.

>>> import expandas as expd
>>> import sklearn.datasets as datasets

>>> df = expd.ModelFrame(datasets.load_iris())
>>> df.head()
   .target  sepal length (cm)  sepal width (cm)  petal length (cm)  \
0        0                5.1               3.5                1.4
1        0                4.9               3.0                1.4
2        0                4.7               3.2                1.3
3        0                4.6               3.1                1.5
4        0                5.0               3.6                1.4

   petal width (cm)
0               0.2
1               0.2
2               0.2
3               0.2
4               0.2

# make columns be readable
>>> df.columns = ['.target', 'sepal length', 'sepal width', 'petal length', 'petal width']

ModelFrame has accessor methods which makes easier access to scikit-learn namespace.

>>> df.cluster.KMeans
<class 'sklearn.cluster.k_means_.KMeans'>

Following table shows scikit-learn module and corresponding ModelFrame module. Some accessors has its abbreviated versions.

Note

Currently, ModelFrame can handle target which consists from a single column. Modules which uses multiple target columns cannot be handled automatically, and marked with (WIP).

`scikit-learn`	`ModelFrame` accessor
`sklearn.cluster`	`ModelFrame.cluster`
`sklearn.covariance`	`ModelFrame.covariance`
`sklearn.cross_validation`	`ModelFrame.cross_validation`, `crv`
`sklearn.decomposition`	`ModelFrame.decomposition`
`sklearn.datasets`	(not accesible from accessor)
`sklearn.dummy`	`ModelFrame.dummy`
`sklearn.ensemble`	`ModelFrame.ensemble`
`sklearn.feature_extraction`	`ModelFrame.feature_extraction`
`sklearn.gaussian_process`	`ModelFrame.gaussian_process` (WIP)
`sklearn.grid_search`	`ModelFrame.grid_search`
`sklearn.isotonic`	`ModelFrame.isotonic`
`sklearn.kernel_approximation`	`ModelFrame.kernel_approximation`
`sklearn.lda`	`ModelFrame.lda`
`sklearn.linear_model`	`ModelFrame.linear_model`
`sklearn.manifold`	`ModelFrame.manifold`
`sklearn.metrics`	`ModelFrame.metrics`
`sklearn.mixture`	`ModelFrame.mixture`
`sklearn.multiclass`	`ModelFrame.multiclass`
`sklearn.naive_bayes`	`ModelFrame.naive_bayes`
`sklearn.neighbors`	`ModelFrame.neighbors`
`sklearn.cross_decomposition`	`ModelFrame.cross_decomposition` (WIP)
`sklearn.pipeline`	`ModelFrame.pipeline`
`sklearn.preprocessing`	`ModelFrame.preprocessing`, `pp`
`sklearn.qda`	`ModelFrame.qda`
`sklearn.semi_supervised`	`ModelFrame.semi_supervised`
`sklearn.svm`	`ModelFrame.svm`
`sklearn.tree`	`ModelFrame.tree`
`sklearn.utils`	(not accesible from accessor)

Thus, you can instanciate each estimator via ModelFrame accessors. Once create an estimator, you can pass it to ModelFrame.fit then predict. ModelFrame automatically uses its data and target properties for each operations.

>>> estimator = df.cluster.KMeans(n_clusters=3)
>>> df.fit(estimator)

>>> predicted = df.predict(estimator)
>>> predicted
0    1
1    1
2    1
...
147    2
148    2
149    0
Length: 150, dtype: int32

ModelFrame preserves the most recently used estimator in estimator atribute, and predicted results in predicted attibute.

>>> df.estimator
KMeans(copy_x=True, init='k-means++', max_iter=300, n_clusters=3, n_init=10,
    n_jobs=1, precompute_distances=True, random_state=None, tol=0.0001,
    verbose=0)

>>> df.predicted
0    1
1    1
2    1
...
147    2
148    2
149    0
Length: 150, dtype: int32

ModelFrame has following methods corresponding to various scikit-learn estimators. The last results are saved as corresponding ModelFrame properties.

`ModelFrame` method	`ModelFrame` property
`ModelFrame.fit`	(None)
`ModelFrame.transform`	(None)
`ModelFrame.fit_transform`	(None)
`ModelFrame.inverse_transform`	(None)
`ModelFrame.predict`	`ModelFrame.predicted`
`ModelFrame.fit_predict`	`ModelFrame.predicted`
`ModelFrame.score`	(None)
`ModelFrame.predict_proba`	`ModelFrame.proba`
`ModelFrame.predict_log_proba`	`ModelFrame.log_proba`
`ModelFrame.decision_function`	`ModelFrame.decision`

Note

If you access to a property before calling ModelFrame methods, ModelFrame automatically calls corresponding method of the latest estimator and return the result.

Following example shows to perform PCA, then revert principal components back to original space.

>>> estimator = df.decomposition.PCA()
>>> df.fit(estimator)

>>> transformed = df.transform(estimator)
>>> transformed.head()
   .target         0         1         2         3
0        0 -2.684207 -0.326607  0.021512  0.001006
1        0 -2.715391  0.169557  0.203521  0.099602
2        0 -2.889820  0.137346 -0.024709  0.019305
3        0 -2.746437  0.311124 -0.037672 -0.075955
4        0 -2.728593 -0.333925 -0.096230 -0.063129

>>> type(transformed)
<class 'expandas.core.frame.ModelFrame'>

>>> transformed.inverse_transform(estimator)
     .target    0    1    2    3
0          0  5.1  3.5  1.4  0.2
1          0  4.9  3.0  1.4  0.2
2          0  4.7  3.2  1.3  0.2
3          0  4.6  3.1  1.5  0.2
4          0  5.0  3.6  1.4  0.2
..       ...  ...  ...  ...  ...
145        2  6.7  3.0  5.2  2.3
146        2  6.3  2.5  5.0  1.9
147        2  6.5  3.0  5.2  2.0
148        2  6.2  3.4  5.4  2.3
149        2  5.9  3.0  5.1  1.8

[150 rows x 5 columns]

Note

columns information will be lost once transformed to principal components.

If ModelFrame both has target and predicted values, the model evaluation can be performed using functions available in ModelFrame.metrics.

>>> estimator = df.svm.SVC()
>>> df.fit(estimator)

>>> df.predict(estimator)
0    0
1    0
2    0
...
147    2
148    2
149    2
Length: 150, dtype: int64

>>> df.predicted
0    0
1    0
2    0
...
147    2
148    2
149    2
Length: 150, dtype: int64

>>> df.metrics.confusion_matrix()
Predicted   0   1   2
Target
0          50   0   0
1           0  48   2
2           0   0  50

Use module level functions¶

Some scikit-learn modules define functions which handle data without instanciating estimators. You can call these functions from accessor methods directly, and ModelFrame will pass corresponding data on background. Following example shows to use sklearn.cluster.k_means function to perform K-means.

Important

When you use module level function, ModelFrame.predicted WILL NOT be updated. Thus, using estimator is recommended.

# no need to pass data explicitly
# sklearn.cluster.kmeans returns centroids, cluster labels and inertia
>>> c, l, i = df.cluster.k_means(n_clusters=3)
>>> l
0     1
1     1
2     1
...
147    2
148    2
149    0
Length: 150, dtype: int32

Pipeline¶

ModelFrame can handle pipeline as the same as normal estimators.

>>> estimators = [('reduce_dim', df.decomposition.PCA()),
...               ('svm', df.svm.SVC())]
>>> pipe = df.pipeline.Pipeline(estimators)
>>> df.fit(pipe)

>>> df.predict(pipe)
0    0
1    0
2    0
...
147    2
148    2
149    2
Length: 150, dtype: int64

Above expression is the same as below:

>>> df2 = df.copy()
>>> df2 = df2.fit_transform(df2.decomposition.PCA())
>>> svm = df2.svm.SVC()
>>> df2.fit(svm)
SVC(C=1.0, cache_size=200, class_weight=None, coef0=0.0, degree=3, gamma=0.0,
  kernel='rbf', max_iter=-1, probability=False, random_state=None,
  shrinking=True, tol=0.001, verbose=False)
>>> df2.predict(svm)
0     0
1     0
2     0
...
147    2
148    2
149    2
Length: 150, dtype: int64

Cross Validation¶

scikit-learn has some classes for cross validation. cross_validation.train_test_split splits data to training and test set. You can access to the function via cross_validation accessor.

>>> train_df, test_df = df.cross_validation.train_test_split()
>>> train_df
     .target  sepal length  sepal width  petal length  petal width
0          0           4.8          3.4           1.9          0.2
1          1           6.3          3.3           4.7          1.6
2          0           4.8          3.4           1.6          0.2
3          2           7.7          2.6           6.9          2.3
4          0           5.4          3.4           1.7          0.2
..       ...           ...          ...           ...          ...
107        0           5.1          3.7           1.5          0.4
108        1           6.7          3.1           4.7          1.5
109        0           4.7          3.2           1.3          0.2
110        0           5.8          4.0           1.2          0.2
111        0           5.1          3.5           1.4          0.2

[112 rows x 5 columns]

>>> test_df
    .target  sepal length  sepal width  petal length  petal width
0         2           6.3          2.7           4.9          1.8
1         0           4.5          2.3           1.3          0.3
2         2           5.8          2.8           5.1          2.4
3         0           4.3          3.0           1.1          0.1
4         0           5.0          3.0           1.6          0.2
..      ...           ...          ...           ...          ...
33        1           6.7          3.1           4.4          1.4
34        0           4.6          3.6           1.0          0.2
35        1           5.7          3.0           4.2          1.2
36        1           5.9          3.0           4.2          1.5
37        2           6.4          2.8           5.6          2.1

[38 rows x 5 columns]

Also, there are some iterative classes which returns indexes for training sets and test sets. You can slice ModelFrame using these indexes.

>>> kf = df.cross_validation.KFold(n=150, n_folds=3)
>>> for train_index, test_index in kf:
...    print('training set shape: ', df.iloc[train_index, :].shape,
...          'test set shape: ', df.iloc[test_index, :].shape)
('training set shape: ', (100, 5), 'test set shape: ', (50, 5))
('training set shape: ', (100, 5), 'test set shape: ', (50, 5))
('training set shape: ', (100, 5), 'test set shape: ', (50, 5))

For further simplification, ModelFrame.cross_validation.iterate can accept such iterators and returns ModelFrame corresponding to training and test data.

>>> kf = df.cross_validation.KFold(n=150, n_folds=3)
>>> for train_df, test_df in df.cross_validation.iterate(kf):
...    print('training set shape: ', train_df.shape,
...          'test set shape: ', test_df.shape)
('training set shape: ', (100, 5), 'test set shape: ', (50, 5))
('training set shape: ', (100, 5), 'test set shape: ', (50, 5))
('training set shape: ', (100, 5), 'test set shape: ', (50, 5))

Grid Search¶

You can perform grid search using ModelFrame.fit.

>>> tuned_parameters = [{'kernel': ['rbf'], 'gamma': [1e-3, 1e-4],
...                     'C': [1, 10, 100]},
...                    {'kernel': ['linear'], 'C': [1, 10, 100]}]

>>> df = expd.ModelFrame(datasets.load_digits())
>>> cv = df.grid_search.GridSearchCV(df.svm.SVC(C=1), tuned_parameters,
...                                  cv=5, scoring='precision')

>>> df.fit(cv)

>>> cv.best_estimator_
SVC(C=10, cache_size=200, class_weight=None, coef0=0.0, degree=3, gamma=0.001,
  kernel='rbf', max_iter=-1, probability=False, random_state=None,
  shrinking=True, tol=0.001, verbose=False)

In addition, ModelFrame.grid_search has a describe function to organize each grid search result as pd.DataFrame accepting estimator.

>>> df.grid_search.describe(cv)
       mean       std    C   gamma  kernel
0.974108  0.013139    1  0.0010     rbf
0.951416  0.020010    1  0.0001     rbf
0.975372  0.011280   10  0.0010     rbf
0.962534  0.020218   10  0.0001     rbf
0.975372  0.011280  100  0.0010     rbf
0.964695  0.016686  100  0.0001     rbf
0.951811  0.018410    1     NaN  linear
0.951811  0.018410   10     NaN  linear
0.951811  0.018410  100     NaN  linear

API:

expandas.core package¶

Submodules¶

expandas.core.accessor module¶

class expandas.core.accessor.AccessorMethods(df, module_name=None, attrs=None)¶

Bases: object

Accessor to related functionalities.

expandas.core.frame module¶

class expandas.core.frame.ModelFrame(data, target=None, *args, **kwargs)¶

Bases: pandas.core.frame.DataFrame

Data structure subclassing pandas.DataFrame to define a metadata to specify target (response variable) and data (explanatory variable / features).

Parameters:

data : same as pandas.DataFrame

target : str or array-like

Column name or values to be used as target

args : arguments passed to pandas.DataFrame

kwargs : keyword arguments passed to pandas.DataFrame

Attributes

`T`	Transpose index and columns
`at`
`axes`
`blocks`	Internal property, property synonym for as_blocks()
`crv`	Property to access `sklearn.cross_validation`
`data`	Return data (explanatory variable / features)
`decision`	Return current estimator’s decision function
`dtypes`	Return the dtypes in this object
`empty`	True if NDFrame is entirely empty [no items]
`estimator`	Return most recently used estimator
`ftypes`	Return the ftypes (indication of sparse/dense and dtype) in this object.
`iat`
`iloc`
`ix`
`loc`
`log_proba`	Return current estimator’s log probabilities
`ndim`	Number of axes / array dimensions
`pp`	Property to access `sklearn.preprocessing`
`predicted`	Return current estimator’s predicted results
`proba`	Return current estimator’s probabilities
`shape`
`size`	number of elements in the NDFrame
`target`	Return target (response variable)
`target_name`	Return target column name
`values`	Numpy representation of NDFrame

cluster
covariance
cross_decomposition
cross_validation
decomposition
dummy
ensemble
feature_extraction
feature_selection
gaussian_process
grid_search
is_copy
isotonic
kernel_approximation
lda
linear_model
manifold
metrics
mixture
multiclass
naive_bayes
neighbors
pipeline
preprocessing
qda
semi_supervised
svm
tree

Methods

`abs`()	Return an object with absolute value taken.
`add`(other[, axis, level, fill_value])	Binary operator add with support to substitute a fill_value for missing data in
`add_prefix`(prefix)	Concatenate prefix string with panel items names.
`add_suffix`(suffix)	Concatenate suffix string with panel items names
`align`(other[, join, axis, level, copy, ...])	Align two object on their axes with the
`all`([axis, bool_only, skipna, level])	Return whether all elements are True over requested axis
`any`([axis, bool_only, skipna, level])	Return whether any element is True over requested axis
`append`(other[, ignore_index, verify_integrity])	Append columns of other to end of this frame’s columns and index, returning a new object.
`apply`(func[, axis, broadcast, raw, reduce, args])	Applies function along input axis of DataFrame.
`applymap`(func)	Apply a function to a DataFrame that is intended to operate elementwise, i.e.
`as_blocks`()	Convert the frame to a dict of dtype -> Constructor Types that each has a homogeneous dtype.
`as_matrix`([columns])	Convert the frame to its Numpy-array representation.
`asfreq`(freq[, method, how, normalize])	Convert all TimeSeries inside to specified frequency using DateOffset objects.
`astype`(dtype[, copy, raise_on_error])	Cast object to input numpy.dtype
`at_time`(time[, asof])	Select values at particular time of day (e.g.
`between_time`(start_time, end_time[, ...])	Select values between particular times of the day (e.g., 9:00-9:30 AM)
`bfill`([axis, inplace, limit, downcast])	Synonym for NDFrame.fillna(method=’bfill’)
`bool`()	Return the bool of a single element PandasObject
`boxplot`([column, by, ax, fontsize, rot, ...])	Make a box plot from DataFrame column optionally grouped by some columns or
`clip`([lower, upper, out])	Trim values at input threshold(s)
`clip_lower`(threshold)	Return copy of the input with values below given value truncated
`clip_upper`(threshold)	Return copy of input with values above given value truncated
`combine`(other, func[, fill_value, overwrite])	Add two DataFrame objects and do not propagate NaN values, so if for a
`combineAdd`(other)	Add two DataFrame objects and do not propagate
`combineMult`(other)	Multiply two DataFrame objects and do not propagate NaN values, so if
`combine_first`(other)	Combine two DataFrame objects and default to non-null values in frame calling the method.
`compound`([axis, skipna, level])	Return the compound percentage of the values for the requested axis
`consolidate`([inplace])	Compute NDFrame with “consolidated” internals (data of each dtype grouped together in a single ndarray).
`convert_objects`([convert_dates, ...])	Attempt to infer better dtype for object columns
`copy`([deep])	Make a copy of this object
`corr`([method, min_periods])	Compute pairwise correlation of columns, excluding NA/null values
`corrwith`(other[, axis, drop])	Compute pairwise correlation between rows or columns of two DataFrame objects.
`count`([axis, level, numeric_only])	Return Series with number of non-NA/null observations over requested axis.
`cov`([min_periods])	Compute pairwise covariance of columns, excluding NA/null values
`cummax`([axis, dtype, out, skipna])	Return cumulative max over requested axis.
`cummin`([axis, dtype, out, skipna])	Return cumulative min over requested axis.
`cumprod`([axis, dtype, out, skipna])	Return cumulative prod over requested axis.
`cumsum`([axis, dtype, out, skipna])	Return cumulative sum over requested axis.
`decision_function`(estimator, args, *kwargs)	Call estimator’s decision_function method.
`describe`([percentile_width, percentiles, ...])	Generate various summary statistics, excluding NaN values.
`diff`([periods])	1st discrete difference of object
`div`(other[, axis, level, fill_value])	Binary operator truediv with support to substitute a fill_value for missing data in
`divide`(other[, axis, level, fill_value])	Binary operator truediv with support to substitute a fill_value for missing data in
`dot`(other)	Matrix multiplication with DataFrame or Series objects
`drop`(labels[, axis, level, inplace])	Return new object with labels in requested axis removed
`drop_duplicates`(args, *kwargs)	Return DataFrame with duplicate rows removed, optionally only
`dropna`([axis, how, thresh, subset, inplace])	Return object with labels on given axis omitted where alternately any
`duplicated`(args, *kwargs)	Return boolean Series denoting duplicate rows, optionally only
`eq`(other[, axis, level])	Wrapper for flexible comparison methods eq
`equals`(other)	Determines if two NDFrame objects contain the same elements.
`eval`(expr, **kwargs)	Evaluate an expression in the context of the calling DataFrame instance.
`ffill`([axis, inplace, limit, downcast])	Synonym for NDFrame.fillna(method=’ffill’)
`fillna`([value, method, axis, inplace, ...])	Fill NA/NaN values using the specified method
`filter`([items, like, regex, axis])	Restrict the info axis to set of items or wildcard
`first`(offset)	Convenience method for subsetting initial periods of time series data
`first_valid_index`()	Return label for first non-NA/null value
`fit`(estimator, args, *kwargs)	Call estimator’s fit method.
`fit_predict`(estimator, args, *kwargs)	Call estimator’s fit_predict method.
`fit_transform`(estimator, args, *kwargs)	Call estimator’s fit_transform method.
`floordiv`(other[, axis, level, fill_value])	Binary operator floordiv with support to substitute a fill_value for missing data in
`from_csv`(path[, header, sep, index_col, ...])	Read delimited file into DataFrame
`from_dict`(data[, orient, dtype])	Construct DataFrame from dict of array-like or dicts
`from_items`(items[, columns, orient])	Convert (key, value) pairs to DataFrame.
`from_records`(data[, index, exclude, ...])	Convert structured or record ndarray to DataFrame
`ge`(other[, axis, level])	Wrapper for flexible comparison methods ge
`get`(key[, default])	Get item from object for given key (DataFrame column, Panel slice, etc.).
`get_dtype_counts`()	Return the counts of dtypes in this object
`get_ftype_counts`()	Return the counts of ftypes in this object
`get_value`(index, col[, takeable])	Quickly retrieve single value at passed column and index
`get_values`()	same as values (but handles sparseness conversions)
`groupby`([by, axis, level, as_index, sort, ...])	Group series using mapper (dict or key function, apply given function
`gt`(other[, axis, level])	Wrapper for flexible comparison methods gt
`has_data`()	Return whether `ModelFrame` has data
`has_target`()	Return whether `ModelFrame` has target
`head`([n])	Returns first n rows
`hist`(data[, column, by, grid, xlabelsize, ...])	Draw histogram of the DataFrame’s series using matplotlib / pylab.
`icol`(i)
`idxmax`([axis, skipna])	Return index of first occurrence of maximum over requested axis.
`idxmin`([axis, skipna])	Return index of first occurrence of minimum over requested axis.
`iget_value`(i, j)
`info`([verbose, buf, max_cols, memory_usage, ...])	Concise summary of a DataFrame.
`insert`(loc, column, value[, allow_duplicates])	Insert column into DataFrame at specified location.
`interpolate`([method, axis, limit, inplace, ...])	Interpolate values according to different methods.
`inverse_transform`(estimator, args, *kwargs)	Call estimator’s inverse_transform method.
`irow`(i[, copy])
`isin`(values)	Return boolean DataFrame showing whether each element in the DataFrame is contained in values.
`isnull`()	Return a boolean same-sized object indicating if the values are null
`iteritems`()	Iterator over (column, series) pairs
`iterkv`(args, *kwargs)	iteritems alias used to get around 2to3. Deprecated
`iterrows`()	Iterate over rows of DataFrame as (index, Series) pairs.
`itertuples`([index])	Iterate over rows of DataFrame as tuples, with index value
`join`(other[, on, how, lsuffix, rsuffix, sort])	Join columns with other DataFrame either on index or on a key column.
`keys`()	Get the ‘info axis’ (see Indexing for more)
`kurt`([axis, skipna, level, numeric_only])	Return unbiased kurtosis over requested axis
`kurtosis`([axis, skipna, level, numeric_only])	Return unbiased kurtosis over requested axis
`last`(offset)	Convenience method for subsetting final periods of time series data
`last_valid_index`()	Return label for last non-NA/null value
`le`(other[, axis, level])	Wrapper for flexible comparison methods le
`load`(path)	Deprecated.
`lookup`(row_labels, col_labels)	Label-based “fancy indexing” function for DataFrame.
`lt`(other[, axis, level])	Wrapper for flexible comparison methods lt
`mad`([axis, skipna, level])	Return the mean absolute deviation of the values for the requested axis
`mask`(cond)	Returns copy whose values are replaced with nan if the
`max`([axis, skipna, level, numeric_only])	This method returns the maximum of the values in the object.
`mean`([axis, skipna, level, numeric_only])	Return the mean of the values for the requested axis
`median`([axis, skipna, level, numeric_only])	Return the median of the values for the requested axis
`memory_usage`([index])	Memory usage of DataFrame columns.
`merge`(right[, how, on, left_on, right_on, ...])	Merge DataFrame objects by performing a database-style join operation by columns or indexes.
`min`([axis, skipna, level, numeric_only])	This method returns the minimum of the values in the object.
`mod`(other[, axis, level, fill_value])	Binary operator mod with support to substitute a fill_value for missing data in
`mode`([axis, numeric_only])	Gets the mode of each element along the axis selected.
`mul`(other[, axis, level, fill_value])	Binary operator mul with support to substitute a fill_value for missing data in
`multiply`(other[, axis, level, fill_value])	Binary operator mul with support to substitute a fill_value for missing data in
`ne`(other[, axis, level])	Wrapper for flexible comparison methods ne
`notnull`()	Return a boolean same-sized object indicating if the values are
`pct_change`([periods, fill_method, limit, freq])	Percent change over given number of periods.
`pivot`([index, columns, values])	Reshape data (produce a “pivot” table) based on column values.
`pivot_table`(args, *kwargs)	Create a spreadsheet-style pivot table as a DataFrame.
`plot`(data[, x, y, kind, ax, subplots, ...])	Make plots of DataFrame using matplotlib / pylab.
`pop`(item)	Return item and drop from frame.
`pow`(other[, axis, level, fill_value])	Binary operator pow with support to substitute a fill_value for missing data in
`predict`(estimator, args, *kwargs)	Call estimator’s predict method.
`predict_log_proba`(estimator, args, *kwargs)	Call estimator’s predict_log_proba method.
`predict_proba`(estimator, args, *kwargs)	Call estimator’s predict_proba method.
`prod`([axis, skipna, level, numeric_only])	Return the product of the values for the requested axis
`product`([axis, skipna, level, numeric_only])	Return the product of the values for the requested axis
`quantile`([q, axis, numeric_only])	Return values at the given quantile over requested axis, a la numpy.percentile.
`query`(expr, **kwargs)	Query the columns of a frame with a boolean expression.
`radd`(other[, axis, level, fill_value])	Binary operator radd with support to substitute a fill_value for missing data in
`rank`([axis, numeric_only, method, ...])	Compute numerical data ranks (1 through n) along axis.
`rdiv`(other[, axis, level, fill_value])	Binary operator rtruediv with support to substitute a fill_value for missing data in
`reindex`([index, columns])	Conform DataFrame to new index with optional filling logic, placing NA/NaN in locations having no value in the previous index.
`reindex_axis`(labels[, axis, method, level, ...])	Conform input object to new index with optional filling logic, placing NA/NaN in locations having no value in the previous index.
`reindex_like`(other[, method, copy, limit])	return an object with matching indicies to myself
`rename`([index, columns])	Alter axes input function or functions.
`rename_axis`(mapper[, axis, copy, inplace])	Alter index and / or columns using input function or functions.
`reorder_levels`(order[, axis])	Rearrange index levels using input order.
`replace`([to_replace, value, inplace, limit, ...])	Replace values given in ‘to_replace’ with ‘value’.
`resample`(rule[, how, axis, fill_method, ...])	Convenience method for frequency conversion and resampling of regular time-series data.
`reset_index`([level, drop, inplace, ...])	For DataFrame with multi-level index, return new DataFrame with labeling information in the columns under the index names, defaulting to ‘level_0’, ‘level_1’, etc.
`rfloordiv`(other[, axis, level, fill_value])	Binary operator rfloordiv with support to substitute a fill_value for missing data in
`rmod`(other[, axis, level, fill_value])	Binary operator rmod with support to substitute a fill_value for missing data in
`rmul`(other[, axis, level, fill_value])	Binary operator rmul with support to substitute a fill_value for missing data in
`rpow`(other[, axis, level, fill_value])	Binary operator rpow with support to substitute a fill_value for missing data in
`rsub`(other[, axis, level, fill_value])	Binary operator rsub with support to substitute a fill_value for missing data in
`rtruediv`(other[, axis, level, fill_value])	Binary operator rtruediv with support to substitute a fill_value for missing data in
`save`(path)	Deprecated.
`score`(estimator, args, *kwargs)	Call estimator’s score method.
`select`(crit[, axis])	Return data corresponding to axis labels matching criteria
`select_dtypes`([include, exclude])	Return a subset of a DataFrame including/excluding columns based on their `dtype`.
`sem`([axis, skipna, level, ddof])	Return unbiased standard error of the mean over requested axis.
`set_axis`(axis, labels)	public verson of axis assignment
`set_index`(keys[, drop, append, inplace, ...])	Set the DataFrame index (row labels) using one or more existing columns.
`set_value`(index, col, value[, takeable])	Put single value at passed column and index
`shift`([periods, freq, axis])	Shift index by desired number of periods with an optional time freq
`skew`([axis, skipna, level, numeric_only])	Return unbiased skew over requested axis
`slice_shift`([periods, axis])	Equivalent to shift without copying data.
`sort`([columns, axis, ascending, inplace, ...])	Sort DataFrame either by labels (along either axis) or by the values in
`sort_index`([axis, by, ascending, inplace, ...])	Sort DataFrame either by labels (along either axis) or by the values in
`sortlevel`([level, axis, ascending, inplace, ...])	Sort multilevel index by chosen axis and primary level.
`squeeze`()	squeeze length 1 dimensions
`stack`([level, dropna])	Pivot a level of the (possibly hierarchical) column labels, returning a DataFrame (or Series in the case of an object with a single level of column labels) having a hierarchical index with a new inner-most level of row labels.
`std`([axis, skipna, level, ddof])	Return unbiased standard deviation over requested axis.
`sub`(other[, axis, level, fill_value])	Binary operator sub with support to substitute a fill_value for missing data in
`subtract`(other[, axis, level, fill_value])	Binary operator sub with support to substitute a fill_value for missing data in
`sum`([axis, skipna, level, numeric_only])	Return the sum of the values for the requested axis
`swapaxes`(axis1, axis2[, copy])	Interchange axes and swap values axes appropriately
`swaplevel`(i, j[, axis])	Swap levels i and j in a MultiIndex on a particular axis
`tail`([n])	Returns last n rows
`take`(indices[, axis, convert, is_copy])	Analogous to ndarray.take
`to_clipboard`([excel, sep])	Attempt to write text representation of object to the system clipboard This can be pasted into Excel, for example.
`to_csv`(args, *kwargs)	Write DataFrame to a comma-separated values (csv) file
`to_dense`()	Return dense representation of NDFrame (as opposed to sparse)
`to_dict`(args, *kwargs)	Convert DataFrame to dictionary.
`to_excel`(args, *kwargs)	Write DataFrame to a excel sheet
`to_gbq`(destination_table[, project_id, ...])	Write a DataFrame to a Google BigQuery table.
`to_hdf`(path_or_buf, key, **kwargs)	activate the HDFStore
`to_html`([buf, columns, col_space, colSpace, ...])	Render a DataFrame as an HTML table.
`to_json`([path_or_buf, orient, date_format, ...])	Convert the object to a JSON string.
`to_latex`([buf, columns, col_space, ...])	Render a DataFrame to a tabular environment table.
`to_msgpack`([path_or_buf])	msgpack (serialize) object to input file path
`to_panel`()	Transform long (stacked) format (DataFrame) into wide (3D, Panel) format.
`to_period`([freq, axis, copy])	Convert DataFrame from DatetimeIndex to PeriodIndex with desired
`to_pickle`(path)	Pickle (serialize) object to input file path
`to_records`([index, convert_datetime64])	Convert DataFrame to record array.
`to_sparse`([fill_value, kind])	Convert to SparseDataFrame
`to_sql`(name, con[, flavor, schema, ...])	Write records stored in a DataFrame to a SQL database.
`to_stata`(fname[, convert_dates, ...])	A class for writing Stata binary dta files from array-like objects
`to_string`([buf, columns, col_space, ...])	Render a DataFrame to a console-friendly tabular output.
`to_timestamp`([freq, how, axis, copy])	Cast to DatetimeIndex of timestamps, at beginning of period
`to_wide`(args, *kwargs)
`transform`(estimator, args, *kwargs)	Call estimator’s transform method.
`transpose`()	Transpose index and columns
`truediv`(other[, axis, level, fill_value])	Binary operator truediv with support to substitute a fill_value for missing data in
`truncate`([before, after, axis, copy])	Truncates a sorted NDFrame before and/or after some particular dates.
`tshift`([periods, freq, axis])	Shift the time index, using the index’s frequency if available
`tz_convert`(tz[, axis, level, copy])	Convert the axis to target time zone.
`tz_localize`(args, *kwargs)	Localize tz-naive TimeSeries to target time zone
`unstack`([level])	Pivot a level of the (necessarily hierarchical) index labels, returning a DataFrame having a new level of column labels whose inner-most level consists of the pivoted index labels.
`update`(other[, join, overwrite, ...])	Modify DataFrame in place using non-NA values from passed DataFrame.
`var`([axis, skipna, level, ddof])	Return unbiased variance over requested axis.
`where`(cond[, other, inplace, axis, level, ...])	Return an object of same shape as self and whose corresponding entries are from self where cond is True and otherwise are from other.
`xs`(key[, axis, level, copy, drop_level])	Returns a cross-section (row(s) or column(s)) from the Series/DataFrame.

cluster = None¶

covariance = None¶

cross_decomposition = None¶

cross_validation = None¶

crv¶: Property to access sklearn.cross_validation

data¶

Return data (explanatory variable / features)

Returns:	data : `ModelFrame`

decision¶

Return current estimator’s decision function

Returns:	probabilities : `ModelFrame`

decision_function(estimator, *args, **kwargs)¶

Call estimator’s decision_function method.

Parameters:

args : arguments passed to decision_function method

kwargs : keyword arguments passed to decision_function method

Returns:

returned : decisions

decomposition = None¶

dummy = None¶

ensemble = None¶

estimator¶

Return most recently used estimator

Returns:	estimator : estimator

feature_extraction = None¶

feature_selection = None¶

fit(estimator, *args, **kwargs)¶

Call estimator’s fit method.

Parameters:

args : arguments passed to fit method

kwargs : keyword arguments passed to fit method

Returns:

returned : None or fitted estimator

fit_predict(estimator, *args, **kwargs)¶

Call estimator’s fit_predict method.

Parameters:

args : arguments passed to fit_predict method

kwargs : keyword arguments passed to fit_predict method

Returns:

returned : predicted result

fit_transform(estimator, *args, **kwargs)¶

Call estimator’s fit_transform method.

Parameters:

args : arguments passed to fit_transform method

kwargs : keyword arguments passed to fit_transform method

Returns:

returned : transformed result

gaussian_process = None¶

grid_search = None¶

has_data()¶

Return whether ModelFrame has data

Returns:	has_data : bool

has_target()¶

Return whether ModelFrame has target

Returns:	has_target : bool

inverse_transform(estimator, *args, **kwargs)¶

Call estimator’s inverse_transform method.

Parameters:

args : arguments passed to inverse_transform method

kwargs : keyword arguments passed to inverse_transform method

Returns:

returned : transformed result

isotonic = None¶

kernel_approximation = None¶

lda = None¶

linear_model = None¶

log_proba¶

Return current estimator’s log probabilities

Returns:	probabilities : `ModelFrame`

manifold = None¶

metrics = None¶

mixture = None¶

multiclass = None¶

naive_bayes = None¶

neighbors = None¶

pipeline = None¶

pp¶: Property to access sklearn.preprocessing

predict(estimator, *args, **kwargs)¶

Call estimator’s predict method.

Parameters:

args : arguments passed to predict method

kwargs : keyword arguments passed to predict method

Returns:

returned : predicted result

predict_log_proba(estimator, *args, **kwargs)¶

Call estimator’s predict_log_proba method.

Parameters:

args : arguments passed to predict_log_proba method

kwargs : keyword arguments passed to predict_log_proba method

Returns:

returned : probabilities

predict_proba(estimator, *args, **kwargs)¶

Call estimator’s predict_proba method.

Parameters:

args : arguments passed to predict_proba method

kwargs : keyword arguments passed to predict_proba method

Returns:

returned : probabilities

predicted¶

Return current estimator’s predicted results

Returns:	predicted : `ModelSeries`

preprocessing = None¶

proba¶

Return current estimator’s probabilities

Returns:	probabilities : `ModelFrame`

qda = None¶

score(estimator, *args, **kwargs)¶

Call estimator’s score method.

Parameters:

args : arguments passed to score method

kwargs : keyword arguments passed to score method

Returns:

returned : score

semi_supervised = None¶

svm = None¶

target¶

Return target (response variable)

Returns:	target : `ModelSeries`

target_name¶

Return target column name

Returns:	target : object

transform(estimator, *args, **kwargs)¶

Call estimator’s transform method.

Parameters:

args : arguments passed to transform method

kwargs : keyword arguments passed to transform method

Returns:

returned : transformed result

tree = None¶

expandas.core.series module¶

class expandas.core.series.ModelSeries(data=None, index=None, dtype=None, name=None, copy=False, fastpath=False)¶

Bases: pandas.core.series.Series

Wrapper for pandas.Series to support sklearn.preprocessing

Attributes

`T`	return the transpose, which is by definition self
`at`
`axes`
`base`	return the base object if the memory of the underlying data is shared
`blocks`	Internal property, property synonym for as_blocks()
`data`	return the data pointer of the underlying data
`dtype`	return the dtype object of the underlying data
`dtypes`	return the dtype object of the underlying data
`empty`	True if NDFrame is entirely empty [no items]
`flags`	return the ndarray.flags for the underlying data
`ftype`	return if the data is sparse\|dense
`ftypes`	return if the data is sparse\|dense
`iat`
`iloc`
`imag`
`is_time_series`
`itemsize`	return the size of the dtype of the item of the underlying data
`ix`
`loc`
`nbytes`	return the number of bytes in the underlying data
`ndim`	return the number of dimensions of the underlying data, by definition 1
`pp`	Property to access `sklearn.preprocessing`
`real`
`shape`	return a tuple of the shape of the underlying data
`size`	return the number of elements in the underlying data
`strides`	return the strides of the underlying data
`values`	Return Series as ndarray

cat
dt
is_copy
preprocessing
str

Methods

`abs`()	Return an object with absolute value taken.
`add`(other[, level, fill_value, axis])	Binary operator add with support to substitute a fill_value for missing data
`add_prefix`(prefix)	Concatenate prefix string with panel items names.
`add_suffix`(suffix)	Concatenate suffix string with panel items names
`align`(other[, join, axis, level, copy, ...])	Align two object on their axes with the
`all`([axis, bool_only, skipna, level])	Return whether all elements are True over requested axis
`any`([axis, bool_only, skipna, level])	Return whether any element is True over requested axis
`append`(to_append[, verify_integrity])	Concatenate two or more Series.
`apply`(func[, convert_dtype, args])	Invoke function on values of Series.
`argmax`([axis, out, skipna])	Index of first occurrence of maximum of values.
`argmin`([axis, out, skipna])	Index of first occurrence of minimum of values.
`argsort`([axis, kind, order])	Overrides ndarray.argsort.
`as_blocks`()	Convert the frame to a dict of dtype -> Constructor Types that each has a homogeneous dtype.
`as_matrix`([columns])	Convert the frame to its Numpy-array representation.
`asfreq`(freq[, method, how, normalize])	Convert all TimeSeries inside to specified frequency using DateOffset objects.
`asof`(where)	Return last good (non-NaN) value in TimeSeries if value is NaN for requested date.
`astype`(dtype[, copy, raise_on_error])	Cast object to input numpy.dtype
`at_time`(time[, asof])	Select values at particular time of day (e.g.
`autocorr`()	Lag-1 autocorrelation
`between`(left, right[, inclusive])	Return boolean Series equivalent to left <= series <= right.
`between_time`(start_time, end_time[, ...])	Select values between particular times of the day (e.g., 9:00-9:30 AM)
`bfill`([axis, inplace, limit, downcast])	Synonym for NDFrame.fillna(method=’bfill’)
`bool`()	Return the bool of a single element PandasObject
`clip`([lower, upper, out])	Trim values at input threshold(s)
`clip_lower`(threshold)	Return copy of the input with values below given value truncated
`clip_upper`(threshold)	Return copy of input with values above given value truncated
`combine`(other, func[, fill_value])	Perform elementwise binary operation on two Series using given function
`combine_first`(other)	Combine Series values, choosing the calling Series’s values first.
`compound`([axis, skipna, level])	Return the compound percentage of the values for the requested axis
`compress`(condition[, axis, out])	Return selected slices of an array along given axis as a Series
`consolidate`([inplace])	Compute NDFrame with “consolidated” internals (data of each dtype grouped together in a single ndarray).
`convert_objects`([convert_dates, ...])	Attempt to infer better dtype for object columns
`copy`([deep])	Make a copy of this object
`corr`(other[, method, min_periods])	Compute correlation with other Series, excluding missing values
`count`([level])	Return number of non-NA/null observations in the Series
`cov`(other[, min_periods])	Compute covariance with Series, excluding missing values
`cummax`([axis, dtype, out, skipna])	Return cumulative max over requested axis.
`cummin`([axis, dtype, out, skipna])	Return cumulative min over requested axis.
`cumprod`([axis, dtype, out, skipna])	Return cumulative prod over requested axis.
`cumsum`([axis, dtype, out, skipna])	Return cumulative sum over requested axis.
`describe`([percentile_width, percentiles, ...])	Generate various summary statistics, excluding NaN values.
`diff`([periods])	1st discrete difference of object
`div`(other[, level, fill_value, axis])	Binary operator truediv with support to substitute a fill_value for missing data
`divide`(other[, level, fill_value, axis])	Binary operator truediv with support to substitute a fill_value for missing data
`dot`(other)	Matrix multiplication with DataFrame or inner-product with Series
`drop`(labels[, axis, level, inplace])	Return new object with labels in requested axis removed
`drop_duplicates`([take_last, inplace])	Return Series with duplicate values removed
`dropna`([axis, inplace])	Return Series without null values
`duplicated`([take_last])	Return boolean Series denoting duplicate values
`eq`(other)
`equals`(other)	Determines if two NDFrame objects contain the same elements.
`factorize`([sort, na_sentinel])	Encode the object as an enumerated type or categorical variable
`ffill`([axis, inplace, limit, downcast])	Synonym for NDFrame.fillna(method=’ffill’)
`fillna`([value, method, axis, inplace, ...])	Fill NA/NaN values using the specified method
`filter`([items, like, regex, axis])	Restrict the info axis to set of items or wildcard
`first`(offset)	Convenience method for subsetting initial periods of time series data
`first_valid_index`()	Return label for first non-NA/null value
`floordiv`(other[, level, fill_value, axis])	Binary operator floordiv with support to substitute a fill_value for missing data
`from_array`(arr[, index, name, dtype, copy, ...])
`from_csv`(path[, sep, parse_dates, header, ...])	Read delimited file into Series
`ge`(other)
`get`(key[, default])	Get item from object for given key (DataFrame column, Panel slice, etc.).
`get_dtype_counts`()	Return the counts of dtypes in this object
`get_ftype_counts`()	Return the counts of ftypes in this object
`get_value`(label[, takeable])	Quickly retrieve single value at passed index label
`get_values`()	same as values (but handles sparseness conversions); is a view
`groupby`([by, axis, level, as_index, sort, ...])	Group series using mapper (dict or key function, apply given function
`gt`(other)
`hasnans`()	return if I have any nans; enables various perf speedups
`head`([n])	Returns first n rows
`hist`([by, ax, grid, xlabelsize, xrot, ...])	Draw histogram of the input series using matplotlib
`idxmax`([axis, out, skipna])	Index of first occurrence of maximum of values.
`idxmin`([axis, out, skipna])	Index of first occurrence of minimum of values.
`iget`(i[, axis])	Return the i-th value or values in the Series by location
`iget_value`(i[, axis])	Return the i-th value or values in the Series by location
`interpolate`([method, axis, limit, inplace, ...])	Interpolate values according to different methods.
`irow`(i[, axis])	Return the i-th value or values in the Series by location
`isin`(values)	Return a boolean `Series` showing whether each element in the `Series` is exactly contained in the passed sequence of `values`.
`isnull`()	Return a boolean same-sized object indicating if the values are null
`item`()	return the first element of the underlying data as a python scalar
`iteritems`()	Lazily iterate over (index, value) tuples
`iterkv`(args, *kwargs)	iteritems alias used to get around 2to3. Deprecated
`keys`()	Alias for index
`kurt`([axis, skipna, level, numeric_only])	Return unbiased kurtosis over requested axis
`kurtosis`([axis, skipna, level, numeric_only])	Return unbiased kurtosis over requested axis
`last`(offset)	Convenience method for subsetting final periods of time series data
`last_valid_index`()	Return label for last non-NA/null value
`le`(other)
`load`(path)	Deprecated.
`lt`(other)
`mad`([axis, skipna, level])	Return the mean absolute deviation of the values for the requested axis
`map`(arg[, na_action])	Map values of Series using input correspondence (which can be
`mask`(cond)	Returns copy whose values are replaced with nan if the
`max`([axis, skipna, level, numeric_only])	This method returns the maximum of the values in the object.
`mean`([axis, skipna, level, numeric_only])	Return the mean of the values for the requested axis
`median`([axis, skipna, level, numeric_only])	Return the median of the values for the requested axis
`min`([axis, skipna, level, numeric_only])	This method returns the minimum of the values in the object.
`mod`(other[, level, fill_value, axis])	Binary operator mod with support to substitute a fill_value for missing data
`mode`()	Returns the mode(s) of the dataset.
`mul`(other[, level, fill_value, axis])	Binary operator mul with support to substitute a fill_value for missing data
`multiply`(other[, level, fill_value, axis])	Binary operator mul with support to substitute a fill_value for missing data
`ne`(other)
`nlargest`([n, take_last])	Return the largest n elements.
`nonzero`()	Return the indices of the elements that are non-zero
`notnull`()	Return a boolean same-sized object indicating if the values are
`nsmallest`([n, take_last])	Return the smallest n elements.
`nunique`([dropna])	Return number of unique elements in the object.
`order`([na_last, ascending, kind, ...])	Sorts Series object, by value, maintaining index-value link.
`pct_change`([periods, fill_method, limit, freq])	Percent change over given number of periods.
`plot`(data[, kind, ax, figsize, use_index, ...])	Make plots of Series using matplotlib / pylab.
`pop`(item)	Return item and drop from frame.
`pow`(other[, level, fill_value, axis])	Binary operator pow with support to substitute a fill_value for missing data
`prod`([axis, skipna, level, numeric_only])	Return the product of the values for the requested axis
`product`([axis, skipna, level, numeric_only])	Return the product of the values for the requested axis
`ptp`([axis, out])
`put`(args, *kwargs)	return a ndarray with the values put
`quantile`([q])	Return value at the given quantile, a la numpy.percentile.
`radd`(other[, level, fill_value, axis])	Binary operator radd with support to substitute a fill_value for missing data
`rank`([method, na_option, ascending, pct])	Compute data ranks (1 through n).
`ravel`([order])	Return the flattened underlying data as an ndarray
`rdiv`(other[, level, fill_value, axis])	Binary operator rtruediv with support to substitute a fill_value for missing data
`reindex`([index])	Conform Series to new index with optional filling logic, placing NA/NaN in locations having no value in the previous index.
`reindex_axis`(labels[, axis])	for compatibility with higher dims
`reindex_like`(other[, method, copy, limit])	return an object with matching indicies to myself
`rename`([index])	Alter axes input function or functions.
`rename_axis`(mapper[, axis, copy, inplace])	Alter index and / or columns using input function or functions.
`reorder_levels`(order)	Rearrange index levels using input order.
`repeat`(reps)	return a new Series with the values repeated reps times
`replace`([to_replace, value, inplace, limit, ...])	Replace values given in ‘to_replace’ with ‘value’.
`resample`(rule[, how, axis, fill_method, ...])	Convenience method for frequency conversion and resampling of regular time-series data.
`reset_index`([level, drop, name, inplace])	Analogous to the `pandas.DataFrame.reset_index()` function, see docstring there.
`reshape`(args, *kwargs)	return an ndarray with the values shape
`rfloordiv`(other[, level, fill_value, axis])	Binary operator rfloordiv with support to substitute a fill_value for missing data
`rmod`(other[, level, fill_value, axis])	Binary operator rmod with support to substitute a fill_value for missing data
`rmul`(other[, level, fill_value, axis])	Binary operator rmul with support to substitute a fill_value for missing data
`round`([decimals, out])	Return a with each element rounded to the given number of decimals.
`rpow`(other[, level, fill_value, axis])	Binary operator rpow with support to substitute a fill_value for missing data
`rsub`(other[, level, fill_value, axis])	Binary operator rsub with support to substitute a fill_value for missing data
`rtruediv`(other[, level, fill_value, axis])	Binary operator rtruediv with support to substitute a fill_value for missing data
`save`(path)	Deprecated.
`searchsorted`(v[, side, sorter])	Find indices where elements should be inserted to maintain order.
`select`(crit[, axis])	Return data corresponding to axis labels matching criteria
`sem`([axis, skipna, level, ddof])	Return unbiased standard error of the mean over requested axis.
`set_axis`(axis, labels)	public verson of axis assignment
`set_value`(label, value[, takeable])	Quickly set single value at passed label.
`shift`([periods, freq, axis])	Shift index by desired number of periods with an optional time freq
`skew`([axis, skipna, level, numeric_only])	Return unbiased skew over requested axis
`slice_shift`([periods, axis])	Equivalent to shift without copying data.
`sort`([axis, ascending, kind, na_position, ...])	Sort values and index labels by value.
`sort_index`([ascending])	Sort object by labels (along an axis)
`sortlevel`([level, ascending, sort_remaining])	Sort Series with MultiIndex by chosen level.
`squeeze`()	squeeze length 1 dimensions
`std`([axis, skipna, level, ddof])	Return unbiased standard deviation over requested axis.
`sub`(other[, level, fill_value, axis])	Binary operator sub with support to substitute a fill_value for missing data
`subtract`(other[, level, fill_value, axis])	Binary operator sub with support to substitute a fill_value for missing data
`sum`([axis, skipna, level, numeric_only])	Return the sum of the values for the requested axis
`swapaxes`(axis1, axis2[, copy])	Interchange axes and swap values axes appropriately
`swaplevel`(i, j[, copy])	Swap levels i and j in a MultiIndex
`tail`([n])	Returns last n rows
`take`(indices[, axis, convert, is_copy])	return Series corresponding to requested indices
`to_clipboard`([excel, sep])	Attempt to write text representation of object to the system clipboard This can be pasted into Excel, for example.
`to_csv`(path[, index, sep, na_rep, ...])	Write Series to a comma-separated values (csv) file
`to_dense`()	Return dense representation of NDFrame (as opposed to sparse)
`to_dict`()	Convert Series to {label -> value} dict
`to_frame`([name])	Convert Series to DataFrame
`to_hdf`(path_or_buf, key, **kwargs)	activate the HDFStore
`to_json`([path_or_buf, orient, date_format, ...])	Convert the object to a JSON string.
`to_msgpack`([path_or_buf])	msgpack (serialize) object to input file path
`to_period`([freq, copy])	Convert TimeSeries from DatetimeIndex to PeriodIndex with desired
`to_pickle`(path)	Pickle (serialize) object to input file path
`to_sparse`([kind, fill_value])	Convert Series to SparseSeries
`to_sql`(name, con[, flavor, schema, ...])	Write records stored in a DataFrame to a SQL database.
`to_string`([buf, na_rep, float_format, ...])	Render a string representation of the Series
`to_timestamp`([freq, how, copy])	Cast to datetimeindex of timestamps, at beginning of period
`tolist`()	Convert Series to a nested list
`transpose`()	return the transpose, which is by definition self
`truediv`(other[, level, fill_value, axis])	Binary operator truediv with support to substitute a fill_value for missing data
`truncate`([before, after, axis, copy])	Truncates a sorted NDFrame before and/or after some particular dates.
`tshift`([periods, freq, axis])	Shift the time index, using the index’s frequency if available
`tz_convert`(tz[, axis, level, copy])	Convert the axis to target time zone.
`tz_localize`(args, *kwargs)	Localize tz-naive TimeSeries to target time zone
`unique`()	Return array of unique values in the object.
`unstack`([level])	Unstack, a.k.a.
`update`(other)	Modify Series in place using non-NA values from passed Series.
`valid`([inplace])
`value_counts`([normalize, sort, ascending, ...])	Returns object containing counts of unique values.
`var`([axis, skipna, level, ddof])	Return unbiased variance over requested axis.
`view`([dtype])
`where`(cond[, other, inplace, axis, level, ...])	Return an object of same shape as self and whose corresponding entries are from self where cond is True and otherwise are from other.
`xs`(key[, axis, level, copy, drop_level])	Returns a cross-section (row(s) or column(s)) from the Series/DataFrame.

pp¶: Property to access sklearn.preprocessing

preprocessing = None¶

to_frame(name=None)¶

Convert Series to DataFrame

Parameters:

name : object, default None

The passed name should substitute for the series name (if it has one).

Returns:

data_frame : DataFrame

Module contents¶

expandas.skaccessors package¶

Subpackages¶

expandas.skaccessors.test package¶

Submodules¶

expandas.skaccessors.test.test_cluster module¶

expandas.skaccessors.test.test_covariance module¶

expandas.skaccessors.test.test_cross_decomposition module¶

expandas.skaccessors.test.test_cross_validation module¶

expandas.skaccessors.test.test_decomposition module¶

expandas.skaccessors.test.test_dummy module¶

expandas.skaccessors.test.test_ensemble module¶

expandas.skaccessors.test.test_feature_extraction module¶

expandas.skaccessors.test.test_feature_selection module¶

expandas.skaccessors.test.test_gaussian_process module¶

expandas.skaccessors.test.test_grid_search module¶

expandas.skaccessors.test.test_isotonic module¶

expandas.skaccessors.test.test_kernel_approximation module¶

expandas.skaccessors.test.test_lda module¶

expandas.skaccessors.test.test_linear_model module¶

expandas.skaccessors.test.test_manifold module¶

expandas.skaccessors.test.test_metrics module¶

expandas.skaccessors.test.test_mixture module¶

expandas.skaccessors.test.test_multiclass module¶

expandas.skaccessors.test.test_naive_bayes module¶

expandas.skaccessors.test.test_neighbors module¶

expandas.skaccessors.test.test_pipeline module¶

expandas.skaccessors.test.test_preprocessing module¶

expandas.skaccessors.test.test_qda module¶

expandas.skaccessors.test.test_semi_supervised module¶

expandas.skaccessors.test.test_svm module¶

expandas.skaccessors.test.test_tree module¶

Module contents¶

Submodules¶

expandas.skaccessors.cluster module¶

class expandas.skaccessors.cluster.ClusterMethods(df, module_name=None, attrs=None)¶

Bases: expandas.core.accessor.AccessorMethods

Accessor to sklearn.cluster.

Attributes

bicluster

Methods

`affinity_propagation`(args, *kwargs)
`dbscan`(args, *kwargs)
`k_means`(n_clusters, args, *kwargs)
`mean_shift`(args, *kwargs)
`spectral_clustering`(args, *kwargs)

affinity_propagation(*args, **kwargs)¶

bicluster = None¶

dbscan(*args, **kwargs)¶

k_means(n_clusters, *args, **kwargs)¶

mean_shift(*args, **kwargs)¶

spectral_clustering(*args, **kwargs)¶

expandas.skaccessors.covariance module¶

class expandas.skaccessors.covariance.CovarianceMethods(df, module_name=None, attrs=None)¶

Bases: expandas.core.accessor.AccessorMethods

Accessor to sklearn.covariance.

Methods

`empirical_covariance`(args, *kwargs)
`ledoit_wolf`(args, *kwargs)
`oas`(args, *kwargs)

empirical_covariance(*args, **kwargs)¶

ledoit_wolf(*args, **kwargs)¶

oas(*args, **kwargs)¶

expandas.skaccessors.cross_validation module¶

class expandas.skaccessors.cross_validation.CrossValidationMethods(df, module_name=None, attrs=None)¶

Bases: expandas.core.accessor.AccessorMethods

Accessor to sklearn.cross_validation.

Methods

`StratifiedShuffleSplit`(args, *kwargs)
`check_cv`(cv, args, *kwargs)
`cross_val_score`(estimator, args, *kwargs)
`iterate`(cv)
`permutation_test_score`(estimator, *args, ...)
`train_test_split`(args, *kwargs)

StratifiedShuffleSplit(*args, **kwargs)¶

check_cv(cv, *args, **kwargs)¶

cross_val_score(estimator, *args, **kwargs)¶

iterate(cv)¶

permutation_test_score(estimator, *args, **kwargs)¶

train_test_split(*args, **kwargs)¶

expandas.skaccessors.decomposition module¶

class expandas.skaccessors.decomposition.DecompositionMethods(df, module_name=None, attrs=None)¶

Bases: expandas.core.accessor.AccessorMethods

Accessor to sklearn.decomposition.

Methods

`dict_learning`(n_components, alpha, *args, ...)
`dict_learning_online`(args, *kwargs)
`fastica`(args, *kwargs)
`sparse_encode`(dictionary, args, *kwargs)

dict_learning(n_components, alpha, *args, **kwargs)¶

dict_learning_online(*args, **kwargs)¶

fastica(*args, **kwargs)¶

sparse_encode(dictionary, *args, **kwargs)¶

expandas.skaccessors.ensemble module¶

class expandas.skaccessors.ensemble.EnsembleMethods(df, module_name=None, attrs=None)¶

Bases: expandas.core.accessor.AccessorMethods

Accessor to sklearn.ensemble.

Attributes

partial_dependence

partial_dependence¶

class expandas.skaccessors.ensemble.PartialDependenceMethods(df, module_name=None, attrs=None)¶

Bases: expandas.core.accessor.AccessorMethods

Methods

`partial_dependence`(gbrt, target_variables, ...)
`plot_partial_dependence`(gbrt, features, **kwargs)

partial_dependence(gbrt, target_variables, **kwargs)¶

plot_partial_dependence(gbrt, features, **kwargs)¶

expandas.skaccessors.feature_extraction module¶

class expandas.skaccessors.feature_extraction.FeatureExtractionMethods(df, module_name=None, attrs=None)¶

Bases: expandas.core.accessor.AccessorMethods

Accessor to sklearn.feature_extraction.

Attributes

image
text

image = None¶

text = None¶

expandas.skaccessors.feature_selection module¶

class expandas.skaccessors.feature_selection.FeatureSelectionMethods(df, module_name=None, attrs=None)¶

Bases: expandas.core.accessor.AccessorMethods

Accessor to sklearn.feature_selection.

expandas.skaccessors.gaussian_process module¶

class expandas.skaccessors.gaussian_process.GaussianProcessMethods(df, module_name=None, attrs=None)¶

Bases: expandas.core.accessor.AccessorMethods

Accessor to sklearn.gaussian_process.

Attributes

`correlation_models`
`regression_models`

correlation_models¶

regression_models¶

class expandas.skaccessors.gaussian_process.RegressionModelsMethods(df, module_name=None, attrs=None)¶: Bases: expandas.core.accessor.AccessorMethods

expandas.skaccessors.grid_search module¶

class expandas.skaccessors.grid_search.GridSearchMethods(df, module_name=None, attrs=None)¶

Bases: expandas.core.accessor.AccessorMethods

Accessor to sklearn.grid_search.

Methods

describe(estimator) Return cross validation results as pd.DataFrame.

describe(estimator)¶: Return cross validation results as pd.DataFrame.

expandas.skaccessors.isotonic module¶

class expandas.skaccessors.isotonic.IsotonicMethods(df, module_name=None, attrs=None)¶

Bases: expandas.core.accessor.AccessorMethods

Accessor to sklearn.isotonic.

Attributes

IsotonicRegression

Methods

`check_increasing`(args, *kwargs)
`isotonic_regression`(args, *kwargs)

IsotonicRegression¶

check_increasing(*args, **kwargs)¶

isotonic_regression(*args, **kwargs)¶

expandas.skaccessors.linear_model module¶

class expandas.skaccessors.linear_model.LinearModelMethods(df, module_name=None, attrs=None)¶

Bases: expandas.core.accessor.AccessorMethods

Accessor to sklearn.linear_model.

Methods

`lars_path`(args, *kwargs)
`lasso_path`(args, *kwargs)
`lasso_stability_path`(args, *kwargs)
`orthogonal_mp_gram`(args, *kwargs)

lars_path(*args, **kwargs)¶

lasso_path(*args, **kwargs)¶

lasso_stability_path(*args, **kwargs)¶

orthogonal_mp_gram(*args, **kwargs)¶

expandas.skaccessors.manifold module¶

class expandas.skaccessors.manifold.ManifoldMethods(df, module_name=None, attrs=None)¶

Bases: expandas.core.accessor.AccessorMethods

Accessor to sklearn.manifold.

Methods

`locally_linear_embedding`(n_neighbors, ...)
`spectral_embedding`(args, *kwargs)

locally_linear_embedding(n_neighbors, n_components, *args, **kwargs)¶

spectral_embedding(*args, **kwargs)¶

expandas.skaccessors.metrics module¶

class expandas.skaccessors.metrics.MetricsMethods(df, module_name=None, attrs=None)¶

Bases: expandas.core.accessor.AccessorMethods

Accessor to sklearn.metrics.

Attributes

pairwise

Methods

`auc`([kind, reorder])
`average_precision_score`(args, *kwargs)
`confusion_matrix`(args, *kwargs)
`consensus_score`(args, *kwargs)
`f1_score`(args, *kwargs)
`fbeta_score`(beta, args, *kwargs)
`hinge_loss`(args, *kwargs)
`log_loss`(args, *kwargs)
`precision_recall_curve`(args, *kwargs)
`precision_recall_fscore_support`(args, *kwargs)
`precision_score`(args, *kwargs)
`recall_score`(args, *kwargs)
`roc_auc_score`(args, *kwargs)
`roc_curve`(args, *kwargs)
`silhouette_samples`(args, *kwargs)
`silhouette_score`(args, *kwargs)

auc(kind='roc', reorder=False, **kwargs)¶

average_precision_score(*args, **kwargs)¶

confusion_matrix(*args, **kwargs)¶

consensus_score(*args, **kwargs)¶

f1_score(*args, **kwargs)¶

fbeta_score(beta, *args, **kwargs)¶

hinge_loss(*args, **kwargs)¶

log_loss(*args, **kwargs)¶

pairwise¶

precision_recall_curve(*args, **kwargs)¶

precision_recall_fscore_support(*args, **kwargs)¶

precision_score(*args, **kwargs)¶

recall_score(*args, **kwargs)¶

roc_auc_score(*args, **kwargs)¶

roc_curve(*args, **kwargs)¶

silhouette_samples(*args, **kwargs)¶

silhouette_score(*args, **kwargs)¶

expandas.skaccessors.multiclass module¶

class expandas.skaccessors.multiclass.MultiClassMethods(df, module_name=None, attrs=None)¶

Bases: expandas.core.accessor.AccessorMethods

Accessor to sklearn.multiclass.

Attributes

`OneVsOneClassifier`
`OneVsRestClassifier`
`OutputCodeClassifier`

Methods

`fit_ecoc`(args, *kwargs)
`fit_ovo`(args, *kwargs)
`fit_ovr`(args, *kwargs)
`predict_ecoc`(args, *kwargs)
`predict_ovo`(args, *kwargs)
`predict_ovr`(args, *kwargs)

OneVsOneClassifier¶

OneVsRestClassifier¶

OutputCodeClassifier¶

fit_ecoc(*args, **kwargs)¶

fit_ovo(*args, **kwargs)¶

fit_ovr(*args, **kwargs)¶

predict_ecoc(*args, **kwargs)¶

predict_ovo(*args, **kwargs)¶

predict_ovr(*args, **kwargs)¶

expandas.skaccessors.neighbors module¶

class expandas.skaccessors.neighbors.NeighborsMethods(df, module_name=None, attrs=None)¶

Bases: expandas.core.accessor.AccessorMethods

Accessor to sklearn.neighbors.

expandas.skaccessors.pipeline module¶

class expandas.skaccessors.pipeline.PipelineMethods(df, module_name=None, attrs=None)¶

Bases: expandas.core.accessor.AccessorMethods

Accessor to sklearn.pipeline.

Attributes

`make_pipeline`
`make_union`

make_pipeline¶

make_union¶

expandas.skaccessors.preprocessing module¶

class expandas.skaccessors.preprocessing.PreprocessingMethods(df, module_name=None, attrs=None)¶

Bases: expandas.core.accessor.AccessorMethods

Accessor to sklearn.preprocessing.

Methods

add_dummy_feature([value])

add_dummy_feature(value=1.0)¶

expandas.skaccessors.svm module¶

class expandas.skaccessors.svm.SVMMethods(df, module_name=None, attrs=None)¶

Bases: expandas.core.accessor.AccessorMethods

Accessor to sklearn.svm.

Attributes

`liblinear`
`libsvm`
`libsvm_sparse`

Methods

l1_min_c(*args, **kwargs)

l1_min_c(*args, **kwargs)¶

liblinear¶

libsvm¶

libsvm_sparse¶