QualitySortPlayerDataTrimAlg
This class is a child of PlayerDataTrimAlg that trims a list of player states based on their expressed performance. It supports a mixed age-quality approach where the residue between two phases is discarded first (this approach is informed by the type in PlayerState). GIMME can benefit from this mixed approach when bootstrapping the system, because simulated data needs to be discarded as real data is added.
QualitySortGridTrimAlg(max_num_model_elements: int, quality_weights: PlayerCharacteristics = {ability = 0.5, engagement = 0.5}, acc_state_residue: boolean = False): void| Name: expected type | Default value | Description |
|---|---|---|
| __quality_weights: PlayerCharacteristics | {ability = 0.5, engagement = 0.5} | These characteristics are used when calculating a state quality. |
| __acc_state_residue: boolean | False | Informs the algorithm if it should consider type (see above). |
consider_state_residue(acc_state_residue: boolean): voidSets the __acc_state_residue flag.
__state_type_filter(elem: PlayerState): booleanCompares states to find if they are real or computed as part of a bootstrap process (see PlayerState).
__calc_quality(state: PlayerState): decimalComputes the quality of a player state, between 0 and 1. A higher value means better quality.
__q_sort(elem: PlayerState): decimalSorting function to order a list by quality.
__calc_quality(state: PlayerState): decimalComputes the quality of a player state, between 0 and 1. A higher value means better quality.
Overrides trimmedList (see PlayerDataTrimAlg).