statistic

Module: missions.operations.statistic

Calculate a two-tailed paired t-test on a result collection for a certain parameter

Furthermore some statistical important values are calculated.

Specification file Parameters

type

Should be statistic

(obligatory, statistic)

metric

list of function values on which we want to make the test

(optional, default: ‘Balanced_accuracy’)

parameter

name of the varying parameter, we want to analyze

(optional, default: ‘__Dataset__’)

filter

dictionary saying which subarray of the csv tabular shall be analyzed

average

parameter, over which one should average

(optional, default: None)

input_collection

Path to the input collection of type ‘result’

related_parameters

list of parameters, being relevant for the related t-test

(optional, default: [“__Dataset__”, Key_Run, Key_Fold])

Exemplary Call

type : statistic
input_path : "result_col_example"
metric : "Balanced_accuracy"
parameter : '__metric__'
related_parameters : ["__Dataset__", "Key_Run", "Key_Fold"]
average : "Key_Run"
filter : {"__metric__":["Balanced_accuracy","k_Balanced_accuracy","soft_Balanced_accuracy"]}

Inheritance diagram for pySPACE.missions.operations.statistic:

Class Summary

StatisticOperation(processes, ...[, ...])	Start only the one StatisticProcess after reading the specification file
StatisticProcess(result_directory, data)	Calculate several statistic metrics on the specified metric and parameter

Classes

StatisticOperation

class pySPACE.missions.operations.statistic.StatisticOperation(processes, operation_spec, result_directory, number_processes, create_process=None)

Bases: pySPACE.missions.operations.base.Operation

Start only the one StatisticProcess after reading the specification file and reducing the performance tabular to the relevant entries

For further calculations, the performance tabular and its metadata.yaml file are copied into the new collection such that other operations can follow, as for example visualization operations.

Class Components Summary

_createProcesses(processes, ...)	Function that creates the process.
consolidate()	Consolidation of the operation’s results
create(operation_spec, result_directory[, ...])	A factory method that creates a statistic operation based on the information given in the operation specification operation_spec.
reduce_tabular(tabular, rel_par, metric, ...)

__init__(processes, operation_spec, result_directory, number_processes, create_process=None)

classmethod create(operation_spec, result_directory, debug=False, input_paths=[])

A factory method that creates a statistic operation based on the information given in the operation specification operation_spec. If debug is TRUE the creation of the statistic processes will not be in a separated thread.

classmethod reduce_tabular(tabular, rel_par, metric, parameter, average)

classmethod _createProcesses(processes, result_directory, data)

Function that creates the process.

Create the Process (it is not distributed over different processes)

consolidate()

Consolidation of the operation’s results

StatisticProcess

class pySPACE.missions.operations.statistic.StatisticProcess(result_directory, data)

Bases: pySPACE.missions.operations.base.Process

Calculate several statistic metrics on the specified metric and parameter

At the moment mean, correlation, difference of means, standard deviation, standard error, p-value, t-value and some basic significance test are calculated and written to a tabular.

Class Components Summary

__call__()	Executes this process on the respective modality
kstest(data)
p_value(p1_list, p2_list)

__init__(result_directory, data)

__call__()

Executes this process on the respective modality

p_value(p1_list, p2_list)

kstest(data)