non-parametric
UK[ˌnɒn pærəˈmetrɪk]US[ˌnɑːn pærəˈmetrɪk]
adj
(of a statistical method or model) not assuming that data belong to any particular distribution or that its structure can be described by a fixed set of parameters.
Morpheme Breakdown
non
para
metric
non
not
para
beside
metric
measure
Etymology
The term is a modern scientific compound, primarily from Greek elements via Latin and French, constructed to describe a specific class of statistical models. The prefix 'non-' (Latin) negates the entire concept, while 'parametric' itself is built from 'para-' (Greek for "beside" or "altered") and 'metric' (Greek for "measure"), originally referring to measurements made alongside or in comparison to a standard. In statistics, "parametric" refers to models based on fixed, measurable parameters. Therefore, 'non-parametric' logically denotes methods that operate outside or independent of such fixed parameter-based assumptions, emphasizing flexibility and distribution-free analysis.
Analysis
Structure: non (not) + para (beside) + metric (measure)
- non: Latin origin, a prefix meaning "not" or "the absence of".
- para: Greek origin (from 'para'), a prefix meaning "beside", "alongside", or "altered".
- metric: Greek origin (from 'metron'), a root meaning "measure".
Examples
The researcher used a non-parametric test because the data did not follow a normal distribution.
Non-parametric methods are often more robust when dealing with outliers or ordinal data.
This machine learning algorithm is non-parametric, meaning it makes no strong assumptions about the form of the underlying function.