A note on the representative adaptive learning algorithm

Michele Berardi; Jaqueson K. Galimberti

doi:10.1016/j.econlet.2014.04.028

A note on the representative adaptive learning algorithm

Michele Berardi, Jaqueson K. Galimberti

Economics

Research output: Contribution to journal › Article › peer-review

Abstract

We compare forecasts from different adaptive learning algorithms and calibrations applied to US real-time data on inflation and growth. We find that the Least Squares with constant gains adjusted to match (past) survey forecasts provides the best overall performance both in terms of forecasting accuracy and in matching (future) survey forecasts. © 2014 Elsevier B.V.

Original language	English
Pages (from-to)	104-107
Number of pages	3
Journal	Economics Letters
Volume	124
Issue number	1
DOIs	https://doi.org/10.1016/j.econlet.2014.04.028
Publication status	Published - 2014

Keywords

Expectations
Forecasting
Learning algorithms
Learning-to-forecast
Least squares
Stochastic gradient

Access to Document

10.1016/j.econlet.2014.04.028

http://www.elsevier.com/homepage/sae/econbase/ecolet/

Cite this

@article{94c4e37a71d34ead843b033554c69bc2,

title = "A note on the representative adaptive learning algorithm",

abstract = "We compare forecasts from different adaptive learning algorithms and calibrations applied to US real-time data on inflation and growth. We find that the Least Squares with constant gains adjusted to match (past) survey forecasts provides the best overall performance both in terms of forecasting accuracy and in matching (future) survey forecasts. {\textcopyright} 2014 Elsevier B.V.",

keywords = "Expectations, Forecasting, Learning algorithms, Learning-to-forecast, Least squares, Stochastic gradient",

author = "Michele Berardi and Galimberti, {Jaqueson K.}",

year = "2014",

doi = "10.1016/j.econlet.2014.04.028",

language = "English",

volume = "124",

pages = "104--107",

journal = "Economics Letters",

issn = "0165-1765",

publisher = "Elsevier BV",

number = "1",

}

TY - JOUR

T1 - A note on the representative adaptive learning algorithm

AU - Berardi, Michele

AU - Galimberti, Jaqueson K.

PY - 2014

Y1 - 2014

N2 - We compare forecasts from different adaptive learning algorithms and calibrations applied to US real-time data on inflation and growth. We find that the Least Squares with constant gains adjusted to match (past) survey forecasts provides the best overall performance both in terms of forecasting accuracy and in matching (future) survey forecasts. © 2014 Elsevier B.V.

AB - We compare forecasts from different adaptive learning algorithms and calibrations applied to US real-time data on inflation and growth. We find that the Least Squares with constant gains adjusted to match (past) survey forecasts provides the best overall performance both in terms of forecasting accuracy and in matching (future) survey forecasts. © 2014 Elsevier B.V.

KW - Expectations

KW - Forecasting

KW - Learning algorithms

KW - Learning-to-forecast

KW - Least squares

KW - Stochastic gradient

U2 - 10.1016/j.econlet.2014.04.028

DO - 10.1016/j.econlet.2014.04.028

M3 - Article

SN - 0165-1765

VL - 124

SP - 104

EP - 107

JO - Economics Letters

JF - Economics Letters

IS - 1

ER -

A note on the representative adaptive learning algorithm

Abstract

Keywords

Access to Document

Fingerprint

Cite this