Myth 2: The myth of the small validation error goal—Beware of over-fitting

K.H. Esbensen; Paul Geladi; Anders Larsen

doi:10.1255/nirn.1295

Myth 2: The myth of the small validation error goal—Beware of over-fitting

K.H. Esbensen, Paul Geladi, Anders Larsen

Department of Geo-energy and Storage

Research output: Contribution to journal › Article › Research

Abstract

A great moment in a chemometrician’s life is to receive a virgin data set of some sort. Will there be a correlation between the raw spectra and appropriate reference values—will the model work out?

Original language	English
Pages (from-to)	18;21
Number of pages	2
Journal	NIR News
Volume	23
Issue number	2
DOIs	https://doi.org/10.1255/nirn.1295
Publication status	Published - Mar 2012

Programme Area

Programme Area 3: Energy Resources

Access to Document

10.1255/nirn.1295

Cite this

@article{a64cc38ddcae493fbe1fc5757af1538e,

title = "Myth 2: The myth of the small validation error goal—Beware of over-fitting",

abstract = "A great moment in a chemometrician{\textquoteright}s life is to receive a virgin data set of some sort. Will there be a correlation between the raw spectra and appropriate reference values—will the model work out?",

author = "K.H. Esbensen and Paul Geladi and Anders Larsen",

year = "2012",

month = mar,

doi = "10.1255/nirn.1295",

language = "English",

volume = "23",

pages = "18;21",

journal = "NIR News",

issn = "0960-3360",

publisher = "Sage Publications",

number = "2",

}

TY - JOUR

T1 - Myth 2: The myth of the small validation error goal—Beware of over-fitting

AU - Esbensen, K.H.

AU - Geladi, Paul

AU - Larsen, Anders

PY - 2012/3

Y1 - 2012/3

N2 - A great moment in a chemometrician’s life is to receive a virgin data set of some sort. Will there be a correlation between the raw spectra and appropriate reference values—will the model work out?

AB - A great moment in a chemometrician’s life is to receive a virgin data set of some sort. Will there be a correlation between the raw spectra and appropriate reference values—will the model work out?

U2 - 10.1255/nirn.1295

DO - 10.1255/nirn.1295

M3 - Article

SN - 0960-3360

VL - 23

SP - 18;21

JO - NIR News

JF - NIR News

IS - 2

ER -

Myth 2: The myth of the small validation error goal—Beware of over-fitting

Abstract

Programme Area

Access to Document

Fingerprint

Cite this