Terms used across the learning material and model documentation. Follow a reference link for the relevant context.
A
- Acquisition metadata
Information about how a measurement was collected, such as timing, sampling, geometry, or instrument settings.
Read more- Agreement
The closeness of measurements intended to measure the same quantity, evaluated in a specified setting.
Read more- Amplitude spectrum
A representation of component amplitudes against frequency. The transform scaling and one- or two-sided convention must be stated.
Read more- Annotation
Information connecting a feature or identifier to an interpreted meaning, often tied to a particular reference and version.
Read more- App
A focused scientific task within Pop Lab, connecting inputs, preparation, available methods, and outputs.
Read more- Artifact manifest
A list of saved filenames, byte sizes and hashes. Use it to check the identity of the listed files. It does not establish scientific validity or include an original input unless that input is actually listed.
Read more
B
- Baseline scope
The observations chosen as the reference for a calculation. In the anomaly-score example, a selected initial interval supplies the mean and sample standard deviation. Choosing that interval does not establish healthy operation.
Read more- Boundary condition
A constraint on a mathematical model at the boundary of its domain.
Read more
C
- Clock-day coverage
The recorded time support within a local calendar day, compared with that day’s expected elapsed duration. A clock change can make a day differ from 24 hours. Coverage alone does not establish wear, hourly activity or eligibility for a 24-hour rhythm.
Read more- Coordinate system
The convention that gives spatial positions a defined orientation, origin, and units.
Read more
D
- DFT
Discrete Fourier transform: a calculation relating a finite sequence of samples to complex frequency coefficients.
Read more- Digital calibration
A numerical conversion using declared coefficients. The Biosignal example uses output = gain × input + offset, with finite coefficients and nonzero gain. An inverse check confirms a numerical relationship; it does not establish instrument calibration or prove a unit label.
Read more
E
- Export precision
The numerical detail actually retained in a saved field. Extra trailing zeros do not restore digits rounded before export. In the airway example, retain the distinction between formatted summary factors and the separately recorded reference values.
Read more
F
- FFT
Fast Fourier transform: an efficient family of algorithms for evaluating a discrete Fourier transform.
Read more
I
- Information criterion
A model-comparison score that depends on fit and parameter count. The model-selection guide prefers lower AIC, BIC or AICc among compatible candidates on the same alignment and topology. A relative ranking does not establish model adequacy.
Read more- Initial condition
The state assigned to a dynamic model at its starting time.
Read more- Interpolation
A rule for estimating values between sampled locations; the choice matters when resampling images or signals.
Read more
M
- Mask
A spatial selection identifying which pixels or voxels belong to a region or an analysis.
Read more- Method
A specified calculation or analysis procedure, interpreted with its inputs, parameters, and assumptions.
Read more- Model
Within the Pop Lab catalog, a scientific domain grouping related toolkits and apps. In simulation, a mathematical description of a system.
Read more
N
- N50 and L50
Order contig lengths from longest to shortest and accumulate them until at least half the assembly length is reached. N50 is the length at that boundary; L50 is the minimum contig count needed to reach it. These summaries alone do not establish assembly correctness.
Read more- Nontrivial split
In the unrooted tree-comparison example, a division of the taxa into two unordered groups, with at least two taxa on each side. Each distinct split is counted once. Preserve the taxon identities when comparing split sets.
Read more- Normalization
A defined transformation or scaling of values. Its meaning depends on the measurement and the analysis.
Read more
P
- Parameter
A value that specifies a model or method, such as a time constant, window length, or fitting option.
Read more- Provenance
The origin and processing history needed to understand where a dataset or output came from.
Read more
R
- Region of interest
A selected spatial region over which an image measurement or analysis is defined.
Read more- Reliability
Consistency of measurements in a specified measurement design; its interpretation depends on the chosen definition and setting.
Read more- Robinson–Foulds distance
For the unrooted routes in the guide, the number of nontrivial splits present in exactly one of two trees with matching taxa. The normalized route uses the stated taxon-count denominator. Zero distance establishes agreement of these split sets, not resolution or biological truth.
Read more
S
- Sampling frequency
The number of regularly spaced samples acquired per second, usually expressed in hertz.
Read more- Source scale
In the multimodal guide, registered or resampled input values kept apart from display-normalized copies. Preserving this scale does not mean that resampling changed nothing, that the file matches the original acquisition bytes or that its units are physically calibrated.
Read more- Spectrogram
A time–frequency display formed from spectra of successive windowed portions of a signal.
Read more- Synthetic data
Values constructed for a known example or simulation rather than collected from the real-world process under study.
Read more
T
- Time constant
A parameter describing the time scale of a dynamic response. In the first-order example, one time constant corresponds to about 63.2% of a step response.
Read more- Toolkit
A grouping of related applications within a Pop Lab scientific model.
Read more- Trial window
A time interval used to align and count responses. In the Neural Coding example, selected trials share a common window around the chosen event, and both counting endpoints are included. Preserve the time reference and inclusion rule; a longer trial need not contribute all its time.
Read more
U
- Unavailable result
A quantity reported as missing, unsupported or mathematically undefined, with its reason retained. It is not a measured zero. The graph example distinguishes a completed run with undefined correlation from a Run refused because no input was supplied.
Read more
V
- Validation
Evaluation of a defined workflow against specified evidence and criteria within a stated scope.
Read more- Voxel
A volume element in a sampled three-dimensional image. Its physical size depends on image spacing.
Read more
W
- Window
A finite segment and, where used, a weighting function applied before a local or whole-record signal calculation.
Read more