Search moodle.org's
Developer Documentation

See Release Notes

  • Bug fixes for general core bugs in 3.10.x will end 8 November 2021 (12 months).
  • Bug fixes for security issues in 3.10.x will end 9 May 2022 (18 months).
  • PHP version: minimum PHP 7.2.0 Note: minimum PHP version has increased since Moodle 3.8. PHP 7.3.x and 7.4.x are supported too.

Base time splitting method.

Copyright: 2016 David Monllao {@link http://www.davidmonllao.com}
License: http://www.gnu.org/copyleft/gpl.html GNU GPL v3 or later
File Size: 311 lines (9 kb)
Included or required:0 times
Referenced: 0 times
Includes or requires: 0 files

Defines 1 class


Class: base  - X-Ref

Base time splitting method.

get_id()   X-Ref
Returns the time splitting method id.

return: string

set_analysable(\core_analytics\analysable $analysable)   X-Ref
Assigns the analysable and updates the time ranges according to the analysable start and end dates.

param: \core_analytics\analysable $analysable
return: void

set_modelid(int $modelid)   X-Ref
Assigns the model id to this time-splitting method it case it needs it.

param: int $modelid

get_analysable()   X-Ref
get_analysable

return: \core_analytics\analysable

is_valid_analysable(\core_analytics\analysable $analysable)   X-Ref
Returns whether the course can be processed by this time splitting method or not.

param: \core_analytics\analysable $analysable
return: bool

ready_to_predict($range)   X-Ref
Should we predict this time range now?

param: array $range
return: bool

ready_to_train($range)   X-Ref
Should we use this time range for training?

param: array $range
return: bool

get_all_ranges()   X-Ref
Returns the ranges used by this time splitting method.

return: array

get_training_ranges()   X-Ref
By default all ranges are for training.

return: array

get_distinct_ranges()   X-Ref
Returns the distinct range indexes in this time splitting method.

return: int[]

get_most_recent_prediction_range()   X-Ref
Returns the most recent range that can be used to predict.

This method is only called when calculating predictions.

return: array

get_range_by_index($rangeindex)   X-Ref
Returns range data by its index.

param: int $rangeindex
return: array|false Range data or false if the index is not part of the existing ranges.

append_rangeindex($sampleid, $rangeindex)   X-Ref
Generates a unique sample id (sample in a range index).

param: int $sampleid
param: int $rangeindex
return: string

infer_sample_info($uniquesampleid)   X-Ref
Returns the sample id and the range index from a uniquesampleid.

param: string $uniquesampleid
return: array array($sampleid, $rangeindex)

include_range_info_in_training_data()   X-Ref
Whether to include the range index in the training data or not.

By default, we consider that the different time ranges included in a time splitting method may not be
compatible between them (i.e. the indicators calculated at the end of the course can easily
differ from indicators calculated at the beginning of the course). So we include the range index as
one of the variables that the machine learning backend uses to generate predictions.

If the indicators calculated using the different time ranges available in this time splitting method
are comparable you can overwrite this method to return false.

Note that:
- This is only relevant for models whose predictions are not based on assumptions
(i.e. the ones using a machine learning backend to generate predictions).
- The ranges can only be included in the training data when
we know the final number of ranges the time splitting method will have. E.g.
We can not know the final number of ranges of a 'daily' time splitting method
as we will have one new range every day.
return: bool

cache_indicator_calculations()   X-Ref
Whether to cache or not the indicator calculations.

Indicator calculations are stored to be reused across models. The calculations
are indexed by the calculation start and end time, and these times depend on the
time-splitting method. You should overwrite this method and return false if the time
frames generated by your time-splitting method are unique and / or can hardly be
reused by further models.

return: bool

valid_for_evaluation()   X-Ref
Is this method valid to evaluate prediction models?

return: bool

validate_ranges()   X-Ref
Validates the time splitting method ranges.

return: void