Outliers ratio¶
A legacy objective-video-quality metric reporting the fraction of model predictions lying outside a declared tolerance interval around subjective mean-opinion scores.
Core Idea¶
The usual interval uses a multiple of the subjective-score standard deviation, but standard, confidence convention, dataset aggregation and whether lower is better must be explicit. Each objective score is paired with its subjective score distribution, classified as inside or outside the permitted band and aggregated as an outlier count divided by total cases. The abstraction is therefore identified by a declared carrier, a transformation or constraint over that carrier, and an invariant that tells an analyst whether the named structure is genuinely present.
Scope of Application¶
Outliers ratio belongs to video quality assessment and is useful where the analyst can specify the typed video quality assessment carrier, defining objects and relations, parameters, conventions, evidence, boundary cases, and comparison targets, then evaluate the video dataset and subjective experiment, objective metric and mapping, mean opinion score and dispersion, tolerance multiplier and interval, paired cases, outlier rule, denominator and uncertainty are explicit. The scope is broad within that domain but bounded by the need for the video dataset and subjective experiment, objective metric and mapping, mean opinion score and dispersion, tolerance multiplier and interval, paired cases, outlier rule, denominator and uncertainty are explicit.
Clarity¶
The abstraction clarifies a crowded vocabulary by making the video dataset and subjective experiment, objective metric and mapping, mean opinion score and dispersion, tolerance multiplier and interval, paired cases, outlier rule, denominator and uncertainty are explicit the center of the account. A claim should name the carrier, the governing operation or relation, the applicable assumptions, and the recognition test.
Manages Complexity¶
Without the abstraction, an analyst must reason directly over many local details: the carrier roles, admissibility assumptions, competing conventions, derived invariants, boundary cases, and proof or validation obligations specific to Outliers ratio. Outliers ratio compresses them into the roles in the structural signature. That compression permits comparison across instances without erasing the variables that determine validity. It also exposes which details may be varied safely and which are constitutive.
Abstract Reasoning¶
- Identify the carrier. State what the elements, states, objects, or observations are: the typed video quality assessment carrier, defining objects and relations, parameters, conventions, evidence, boundary cases, and comparison targets. Reject examples whose alleged carrier belongs to a different problem. 2. Lock the constitutive rule. Express the video dataset and subjective experiment, objective metric and mapping, mean opinion score and dispersion, tolerance multiplier and interval, paired cases, outlier rule, denominator and uncertainty are explicit independently of one notation or implementation.
Knowledge Transfer¶
Knowledge transfers strongly among subfields of video quality assessment because they reuse the typed video quality assessment carrier, defining objects and relations, parameters, conventions, evidence, boundary cases, and comparison targets, Each objective score is paired with its subjective score distribution, classified as inside or outside the permitted band and aggregated as an outlier count divided by total cases., and type the carrier, state every parameter and convention in the definition, test that the video dataset and subjective experiment, objective metric and mapping, mean opinion score and dispersion, tolerance multiplier and interval, paired cases, outlier rule, denominator and uncertainty are explicit, compare the nearest accepted identity, and report counterexamples, uncertainty, and limiting cases.
Relationships to Other Abstractions¶
Current abstraction Outliers ratio Domain-specific
Parents (1) — more general patterns this builds on
-
Outliers ratio is a kind of Evaluation Prime
The proposed strict upward parent is
prime:evaluation.
Hierarchy path (1) — routes to 1 parentless root
- Outliers ratio → Evaluation → Comparison → Self Checking
Neighborhood in Abstraction Space¶
Outliers ratio sits in a moderately populated region (43rd percentile for distinctiveness): it has near-neighbors but no dense thicket of look-alikes.
Family — Psychometrics, Testing & Measurement Bias (24 abstractions)
Nearest neighbors
- Peak signal-to-noise ratio — 0.90
- Analytic and enumerative statistical studies — 0.89
- Correspondence analysis — 0.89
- L-estimator — 0.89
- Correlation ratio — 0.89
Computed from structural-signature embeddings · 2026-09-08