Mercurial > pylearn

annotate statscollector.py @ 55:66619ce44497

Find changesets by keywords (author, files, the commit message), revision number or hash, or revset expression.
Efficient implementation of getitem for ArrayDataSet
author Yoshua Bengio <bengioy@iro.umontreal.ca>
date Tue, 29 Apr 2008 15:05:12 -0400
parents 2cd82666b9a7
children f62a03c9d485
rev   line source
1
2cd82666b9a7 Added statscollector and started writing dataset and learner.
bengioy@esprit.iro.umontreal.ca
parents:
diff changeset
1
2cd82666b9a7 Added statscollector and started writing dataset and learner.
bengioy@esprit.iro.umontreal.ca
parents:
diff changeset
2 from numpy import *
2cd82666b9a7 Added statscollector and started writing dataset and learner.
bengioy@esprit.iro.umontreal.ca
parents:
diff changeset
3
2cd82666b9a7 Added statscollector and started writing dataset and learner.
bengioy@esprit.iro.umontreal.ca
parents:
diff changeset
4 class StatsCollector(object):
2cd82666b9a7 Added statscollector and started writing dataset and learner.
bengioy@esprit.iro.umontreal.ca
parents:
diff changeset
5 """A StatsCollector object is used to record performance statistics during training
2cd82666b9a7 Added statscollector and started writing dataset and learner.
bengioy@esprit.iro.umontreal.ca
parents:
diff changeset
6 or testing of a learner. It can be configured to measure different things and
2cd82666b9a7 Added statscollector and started writing dataset and learner.
bengioy@esprit.iro.umontreal.ca
parents:
diff changeset
7 accumulate the appropriate statistics. From these statistics it can be interrogated
2cd82666b9a7 Added statscollector and started writing dataset and learner.
bengioy@esprit.iro.umontreal.ca
parents:
diff changeset
8 to obtain performance measures of interest (such as maxima, minima, mean, standard
2cd82666b9a7 Added statscollector and started writing dataset and learner.
bengioy@esprit.iro.umontreal.ca
parents:
diff changeset
9 deviation, standard error, etc.). Optionally, the observations can be weighted
2cd82666b9a7 Added statscollector and started writing dataset and learner.
bengioy@esprit.iro.umontreal.ca
parents:
diff changeset
10 (yielded weighted mean, weighted variance, etc., where applicable). The statistics
2cd82666b9a7 Added statscollector and started writing dataset and learner.
bengioy@esprit.iro.umontreal.ca
parents:
diff changeset
11 that are desired can be specified among a list supported by the StatsCollector
2cd82666b9a7 Added statscollector and started writing dataset and learner.
bengioy@esprit.iro.umontreal.ca
parents:
diff changeset
12 class or subclass. When some statistics are requested, others become automatically
2cd82666b9a7 Added statscollector and started writing dataset and learner.
bengioy@esprit.iro.umontreal.ca
parents:
diff changeset
13 available (e.g., sum or mean)."""
2cd82666b9a7 Added statscollector and started writing dataset and learner.
bengioy@esprit.iro.umontreal.ca
parents:
diff changeset
14
2cd82666b9a7 Added statscollector and started writing dataset and learner.
bengioy@esprit.iro.umontreal.ca
parents:
diff changeset
15 default_statistics = [mean,standard_deviation,min,max]
2cd82666b9a7 Added statscollector and started writing dataset and learner.
bengioy@esprit.iro.umontreal.ca
parents:
diff changeset
16
2cd82666b9a7 Added statscollector and started writing dataset and learner.
bengioy@esprit.iro.umontreal.ca
parents:
diff changeset
17 __init__(self,n_quantities_observed, statistics=default_statistics):
2cd82666b9a7 Added statscollector and started writing dataset and learner.
bengioy@esprit.iro.umontreal.ca
parents:
diff changeset
18 self.n_quantities_observed=n_quantities_observed
2cd82666b9a7 Added statscollector and started writing dataset and learner.
bengioy@esprit.iro.umontreal.ca
parents:
diff changeset
19
2cd82666b9a7 Added statscollector and started writing dataset and learner.
bengioy@esprit.iro.umontreal.ca
parents:
diff changeset
20 clear(self):
2cd82666b9a7 Added statscollector and started writing dataset and learner.
bengioy@esprit.iro.umontreal.ca
parents:
diff changeset
21 raise NotImplementedError
2cd82666b9a7 Added statscollector and started writing dataset and learner.
bengioy@esprit.iro.umontreal.ca
parents:
diff changeset
22
2cd82666b9a7 Added statscollector and started writing dataset and learner.
bengioy@esprit.iro.umontreal.ca
parents:
diff changeset
23 update(self,observations):
2cd82666b9a7 Added statscollector and started writing dataset and learner.
bengioy@esprit.iro.umontreal.ca
parents:
diff changeset
24 """The observations is a numpy vector of length n_quantities_observed. Some
2cd82666b9a7 Added statscollector and started writing dataset and learner.
bengioy@esprit.iro.umontreal.ca
parents:
diff changeset
25 entries can be 'missing' (with a NaN entry) and will not be counted in the
2cd82666b9a7 Added statscollector and started writing dataset and learner.
bengioy@esprit.iro.umontreal.ca
parents:
diff changeset
26 statistics."""
2cd82666b9a7 Added statscollector and started writing dataset and learner.
bengioy@esprit.iro.umontreal.ca
parents:
diff changeset
27 raise NotImplementedError
2cd82666b9a7 Added statscollector and started writing dataset and learner.
bengioy@esprit.iro.umontreal.ca
parents:
diff changeset
28
2cd82666b9a7 Added statscollector and started writing dataset and learner.
bengioy@esprit.iro.umontreal.ca
parents:
diff changeset
29 __getattr__(self, statistic)
2cd82666b9a7 Added statscollector and started writing dataset and learner.
bengioy@esprit.iro.umontreal.ca
parents:
diff changeset
30 """Return a particular statistic, which may be inferred from the collected statistics.
2cd82666b9a7 Added statscollector and started writing dataset and learner.
bengioy@esprit.iro.umontreal.ca
parents:
diff changeset
31 The argument is a string naming that statistic."""
2cd82666b9a7 Added statscollector and started writing dataset and learner.
bengioy@esprit.iro.umontreal.ca
parents:
diff changeset
32
2cd82666b9a7 Added statscollector and started writing dataset and learner.
bengioy@esprit.iro.umontreal.ca
parents:
diff changeset
33
2cd82666b9a7 Added statscollector and started writing dataset and learner.
bengioy@esprit.iro.umontreal.ca
parents:
diff changeset
34
2cd82666b9a7 Added statscollector and started writing dataset and learner.
bengioy@esprit.iro.umontreal.ca
parents:
diff changeset
35
2cd82666b9a7 Added statscollector and started writing dataset and learner.
bengioy@esprit.iro.umontreal.ca
parents:
diff changeset
36
2cd82666b9a7 Added statscollector and started writing dataset and learner.
bengioy@esprit.iro.umontreal.ca
parents:
diff changeset
37