Learning data discretization via convex optimization

Kód výsledku v IS VaVaI

<a href="https://www.isvavai.cz/riv?ss=detail&h=RIV%2F68407700%3A21230%2F18%3A00318220" target="_blank" >RIV/68407700:21230/18:00318220 - isvavai.cz</a>

Výsledek na webu

<a href="http://link.springer.com/article/10.1007/s10994-017-5654-4" target="_blank" >http://link.springer.com/article/10.1007/s10994-017-5654-4</a>

DOI - Digital Object Identifier

<a href="http://dx.doi.org/10.1007/s10994-017-5654-4" target="_blank" >10.1007/s10994-017-5654-4</a>

Jazyk výsledku

angličtina

Název v původním jazyce

Learning data discretization via convex optimization

Popis výsledku v původním jazyce

Discretization of continuous input functions into piecewise constant or piecewise linear approximations is needed in many mathematical modeling problems. It has been shown that choosing the length of the piecewise segments adaptively based on data samples leads to improved accuracy of the subsequent processing such as classification. Traditional approaches are often tied to a particular classification model which results in local greedy optimization of a criterion function. This paper proposes a technique for learning the discretization parameters along with the parameters of a decision function in a convex optimization of the true objective. The general formulation is applicable to a wide range of learning problems. Empirical evaluation demonstrates that the proposed convex algorithms yield models with fewer number of parameters with comparable or better accuracy than the existing methods.

Název v anglickém jazyce

Learning data discretization via convex optimization

Popis výsledku anglicky

Discretization of continuous input functions into piecewise constant or piecewise linear approximations is needed in many mathematical modeling problems. It has been shown that choosing the length of the piecewise segments adaptively based on data samples leads to improved accuracy of the subsequent processing such as classification. Traditional approaches are often tied to a particular classification model which results in local greedy optimization of a criterion function. This paper proposes a technique for learning the discretization parameters along with the parameters of a decision function in a convex optimization of the true objective. The general formulation is applicable to a wide range of learning problems. Empirical evaluation demonstrates that the proposed convex algorithms yield models with fewer number of parameters with comparable or better accuracy than the existing methods.

Druh

J_imp - Článek v periodiku v databázi Web of Science

CEP obor

—

OECD FORD obor

10201 - Computer sciences, information science, bioinformathics (hardware development to be 2.2, social aspect to be 5.8)

Projekt

<a href="/cs/project/GA16-05872S" target="_blank" >GA16-05872S: Pravděpodobnostní grafové modely a hluboké učení</a><br>

Návaznosti

S - Specificky vyzkum na vysokych skolach

Rok uplatnění

2018

Kód důvěrnosti údajů

S - Úplné a pravdivé údaje o projektu nepodléhají ochraně podle zvláštních právních předpisů

Název periodika

Machine Learning

ISSN

0885-6125

e-ISSN

1573-0565

Svazek periodika

107

Číslo periodika v rámci svazku

2

Stát vydavatele periodika

US - Spojené státy americké

Počet stran výsledku

23

Strana od-do

333-355

Kód UT WoS článku

000423385500002

EID výsledku v databázi Scopus

2-s2.0-85023205105