-
Notifications
You must be signed in to change notification settings - Fork 1k
/
evaluate.py
281 lines (228 loc) · 11 KB
/
evaluate.py
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
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
"""The :mod:`evaluate` module defines the :func:`evaluate` function and
:class:`GridSearch` class """
from __future__ import (absolute_import, division, print_function,
unicode_literals)
from collections import defaultdict
import time
import os
import numpy as np
from six import iteritems
from six import itervalues
from itertools import product
from . import accuracy
from .dump import dump
def evaluate(algo, data, measures=['rmse', 'mae'], with_dump=False,
dump_dir=None, verbose=1):
"""Evaluate the performance of the algorithm on given data.
Depending on the nature of the ``data`` parameter, it may or may not
perform cross validation.
Args:
algo(:obj:`AlgoBase \
<surprise.prediction_algorithms.algo_base.AlgoBase>`):
The algorithm to evaluate.
data(:obj:`Dataset <surprise.dataset.Dataset>`): The dataset on which
to evaluate the algorithm.
measures(list of string): The performance measures to compute. Allowed
names are function names as defined in the :mod:`accuracy
<surprise.accuracy>` module. Default is ``['rmse', 'mae']``.
with_dump(bool): If True, the predictions and the algorithm will be
dumped for later further analysis at each fold (see :ref:`FAQ
<serialize_an_algorithm>`). The file names will be set as:
``'<date>-<algorithm name>-<fold number>'``. Default is ``False``.
dump_dir(str): The directory where to dump to files. Default is
``'~/.surprise_data/dumps/'``.
verbose(int): Level of verbosity. If 0, nothing is printed. If 1
(default), accuracy measures for each folds are printed, with a
final summary. If 2, every prediction is printed.
Returns:
A dictionary containing measures as keys and lists as values. Each list
contains one entry per fold.
"""
performances = CaseInsensitiveDefaultDict(list)
if verbose:
print('Evaluating {0} of algorithm {1}.'.format(
', '.join((m.upper() for m in measures)),
algo.__class__.__name__))
print()
for fold_i, (trainset, testset) in enumerate(data.folds()):
if verbose:
print('-' * 12)
print('Fold ' + str(fold_i + 1))
# train and test algorithm. Keep all rating predictions in a list
algo.train(trainset)
predictions = algo.test(testset, verbose=(verbose == 2))
# compute needed performance statistics
for measure in measures:
f = getattr(accuracy, measure.lower())
performances[measure].append(f(predictions, verbose=verbose))
if with_dump:
if dump_dir is None:
dump_dir = os.path.expanduser('~') + '/.surprise_data/dumps/'
if not os.path.exists(dump_dir):
os.makedirs(dump_dir)
date = time.strftime('%y%m%d-%Hh%Mm%S', time.localtime())
file_name = date + '-' + algo.__class__.__name__
file_name += '-fold{0}'.format(fold_i + 1)
file_name = os.path.join(dump_dir, file_name)
dump(file_name, predictions, trainset, algo)
if verbose:
print('-' * 12)
print('-' * 12)
for measure in measures:
print('Mean {0:4s}: {1:1.4f}'.format(
measure.upper(), np.mean(performances[measure])))
print('-' * 12)
print('-' * 12)
return performances
class GridSearch:
"""The :class:`GridSearch` class, used to evaluate the performance of an
algorithm on various combinations of parameters, and extract the best
combination. It is analogous to `GridSearchCV
<http://scikit-learn.org/stable/modules/generated/sklearn.
model_selection.GridSearchCV.html>`_ from scikit-learn.
See :ref:`User Guide <tuning_algorithm_parameters>` for usage.
Args:
algo_class(:obj:`AlgoBase \
<surprise.prediction_algorithms.algo_base.AlgoBase>`):
A class object of of the algorithm to evaluate.
param_grid (dict):
The dictionary has algo_class parameters as keys (string) and list
of parameters as the desired values to try. All combinations will
be evaluated with desired algorithm.
measures(list of string):
The performance measures to compute. Allowed names are function
names as defined in the :mod:`accuracy <surprise.accuracy>` module.
Default is ``['rmse', 'mae']``.
verbose(int):
Level of verbosity. If ``0``, nothing is printed. If ``1``,
accuracy measures for each parameters combination are printed, with
combination values. If ``2``, folds accuracy values are also
printed. Default is ``1``.
Attributes:
cv_results (dict of arrays):
A dict that contains all parameters and accuracy information for
each combination. Can be imported into a pandas `DataFrame`.
best_estimator (dict of AlgoBase):
Using an accuracy measure as key, get the estimator that gave the
best accuracy results for the chosen measure.
best_score (dict of floats):
Using an accuracy measure as key, get the best score achieved for
that measure.
best_params (dict of dicts):
Using an accuracy measure as key, get the parameters combination
that gave the best accuracy results for the chosen measure.
best_index (dict of ints):
Using an accuracy measure as key, get the index that can be used
with `cv_results_` that achieved the highest accuracy for that
measure.
"""
def __init__(self, algo_class, param_grid, measures=['rmse', 'mae'],
verbose=1):
self.best_params = CaseInsensitiveDefaultDict(list)
self.best_index = CaseInsensitiveDefaultDict(list)
self.best_score = CaseInsensitiveDefaultDict(list)
self.best_estimator = CaseInsensitiveDefaultDict(list)
self.cv_results = defaultdict(list)
self.algo_class = algo_class
self.param_grid = param_grid
self.measures = [measure.upper() for measure in measures]
self.verbose = verbose
# As sim_options and bsl_options are dictionaries, they require a
# special treatment.
if 'sim_options' in param_grid:
sim_options = param_grid['sim_options']
sim_options_list = [dict(zip(sim_options, v)) for v in
product(*sim_options.values())]
param_grid['sim_options'] = sim_options_list
if 'bsl_options' in param_grid:
bsl_options = param_grid['bsl_options']
bsl_options_list = [dict(zip(bsl_options, v)) for v in
product(*bsl_options.values())]
param_grid['bsl_options'] = bsl_options_list
self.param_combinations = [dict(zip(param_grid, v)) for v in
product(*param_grid.values())]
print(self.param_combinations)
def evaluate(self, data):
"""Runs the grid search on dataset.
Class instance attributes can be accessed after the evaluate is done.
Args:
data (:obj:`Dataset <surprise.dataset.Dataset>`): The dataset on
which to evaluate the algorithm.
"""
num_of_combinations = len(self.param_combinations)
params = []
scores = []
# evaluate each combination of parameters using the evaluate method
for combination_index, combination in enumerate(
self.param_combinations):
params.append(combination)
if self.verbose >= 1:
print('-' * 12)
print('Parameters combination {} of {}'.
format(combination_index + 1, num_of_combinations))
print('params: ', combination)
# the algorithm to use along with the combination parameters
algo_instance = self.algo_class(**combination)
evaluate_results = evaluate(algo_instance, data,
measures=self.measures,
verbose=(self.verbose == 2))
# measures as keys and folds average as values
mean_score = {}
for measure in self.measures:
mean_score[measure] = np.mean(evaluate_results[measure])
scores.append(mean_score)
if self.verbose == 1:
print('-' * 12)
for measure in self.measures:
print('Mean {0:4s}: {1:1.4f}'.format(
measure, mean_score[measure]))
print('-' * 12)
# Add all scores and parameters lists to dict
self.cv_results['params'] = params
self.cv_results['scores'] = scores
# Add accuracy measures and algorithm parameters as keys to dict
for param, score in zip(params, scores):
for param_key, score_key in zip(param.keys(), score.keys()):
self.cv_results[param_key].append(param[param_key])
self.cv_results[score_key].append(score[score_key])
# Get the best results
for measure in self.measures:
if measure == 'FCP':
best_dict = max(self.cv_results['scores'],
key=lambda x: x[measure])
else:
best_dict = min(self.cv_results['scores'],
key=lambda x: x[measure])
self.best_score[measure] = best_dict[measure]
self.best_index[measure] = self.cv_results['scores'].index(
best_dict)
self.best_params[measure] = self.cv_results['params'][
self.best_index[measure]]
self.best_estimator[measure] = self.algo_class(
**self.best_params[measure])
class CaseInsensitiveDefaultDict(defaultdict):
"""From here:
http://stackoverflow.com/questions/2082152/case-insensitive-dictionary.
As pointed out in the comments, this only covers a few cases and we
should override a lot of other methods, but oh well...
Used for the returned dict, so that users can use perf['RMSE'] or
perf['rmse'] indifferently.
"""
def __setitem__(self, key, value):
super(CaseInsensitiveDefaultDict, self).__setitem__(key.lower(), value)
def __getitem__(self, key):
return super(CaseInsensitiveDefaultDict, self).__getitem__(key.lower())
def print_perf(performances):
# retrieve number of folds. Kind of ugly...
n_folds = [len(values) for values in itervalues(performances)][0]
row_format = '{:<8}' * (n_folds + 2)
s = row_format.format(
'',
*['Fold {0}'.format(i + 1) for i in range(n_folds)] + ['Mean'])
s += '\n'
s += '\n'.join(row_format.format(
key.upper(),
*['{:1.4f}'.format(v) for v in vals] +
['{:1.4f}'.format(np.mean(vals))])
for (key, vals) in iteritems(performances))
print(s)