Applying wrapper-based variable selection techniques to predict MFIs profitability: evidence from Peru

Fabio Pietrapiana; José M. Feria Domínguez; Alicia Troncoso

doi:10.1080/19439342.2021.1884119

Applying wrapper-based variable selection techniques to predict MFIs profitability: evidence from Peru

Fabio Pietrapiana, José M. Feria Domínguez, Alicia Troncoso

Producción científica: Contribución a una revista › Artículo (Contribución a Revista) › revisión exhaustiva

3 Citas (Scopus)

Resumen

En este trabajo, analizamos los principales factores que explican la rentabilidad (ROA) de las Instituciones de Microfinanzas (IMF) en Perú desde 2011 hasta 2107. Aplicamos tres técnicas de envoltura a una muestra de 168 IMF peruanas y 69 atributos obtenidos de la base de datos MIX Market. Después de ejecutar los algoritmos M5', vecinos knearest (KNN) y Random Forest, encontramos que el algoritmo M5' proporciona el mejor ajuste para predecir el ROA. En particular, la variable clave del árbol de regresión es el porcentaje de gastos sobre activos y, dependiendo de su valor, le sigue la utilidad neta después de impuestos y antes de donaciones o márgenes de utilidad.

Idioma original	Inglés estadounidense
Páginas (desde-hasta)	84-99
Número de páginas	16
Publicación	Default journal
Volumen	13
N.º	1
DOI	https://doi.org/10.1080/19439342.2021.1884119
Estado	Publicada - 2 ene. 2021
Publicado de forma externa	Sí

COAR

Artículo

Categoría OCDE

Ingeniería industrial

Temas Repositorio Ulima

Capitalistas y financieros
Financial services industry
Microcrédito
Microfinance
Profitability
Rentabilidad

Acceder al documento

10.1080/19439342.2021.1884119

Otros archivos y enlaces

Citar esto

@article{fd483470c45c41768bee45a345dffcfd,

title = "Applying wrapper-based variable selection techniques to predict MFIs profitability: evidence from Peru",

abstract = "In this paper, we analyse the main factors explaining the profitability (ROA) of Microfinance Institutions (MFIs) in Peru from 2011 to 2107. We apply three wrapper techniques to asample of 168 Peruvians MFIs and 69 attributes obtained from MIX Market database. After running the algorithms M5ʹ, knearest neighbours (KNN) and Random Forest, we find that the M5ʹ algorithm provides the best fit for predicting ROA. Particularly, the key variable of the regression tree is the percentage of expenses over assets and, depending on its value, it is followed by net income after taxes and before donations, or profit margins.",

keywords = "Microfinance Institutions (MFIs), k Nearest Neighbours (KNN), m5, peru, random Forest (RF), return on Assets (ROA), wrapper Techniques",

author = "Fabio Pietrapiana and {Feria Dom{\'i}nguez}, {Jos{\'e} M.} and Alicia Troncoso",

note = "Publisher Copyright: {\textcopyright} 2021 Informa UK Limited, trading as Taylor & Francis Group.",

year = "2021",

month = jan,

day = "2",

doi = "10.1080/19439342.2021.1884119",

language = "American English",

volume = "13",

pages = "84--99",

journal = "Default journal",

number = "1",

}

TY - JOUR

T1 - Applying wrapper-based variable selection techniques to predict MFIs profitability: evidence from Peru

AU - Pietrapiana, Fabio

AU - Feria Domínguez, José M.

AU - Troncoso, Alicia

PY - 2021/1/2

Y1 - 2021/1/2

N2 - In this paper, we analyse the main factors explaining the profitability (ROA) of Microfinance Institutions (MFIs) in Peru from 2011 to 2107. We apply three wrapper techniques to asample of 168 Peruvians MFIs and 69 attributes obtained from MIX Market database. After running the algorithms M5ʹ, knearest neighbours (KNN) and Random Forest, we find that the M5ʹ algorithm provides the best fit for predicting ROA. Particularly, the key variable of the regression tree is the percentage of expenses over assets and, depending on its value, it is followed by net income after taxes and before donations, or profit margins.

AB - In this paper, we analyse the main factors explaining the profitability (ROA) of Microfinance Institutions (MFIs) in Peru from 2011 to 2107. We apply three wrapper techniques to asample of 168 Peruvians MFIs and 69 attributes obtained from MIX Market database. After running the algorithms M5ʹ, knearest neighbours (KNN) and Random Forest, we find that the M5ʹ algorithm provides the best fit for predicting ROA. Particularly, the key variable of the regression tree is the percentage of expenses over assets and, depending on its value, it is followed by net income after taxes and before donations, or profit margins.

KW - Microfinance Institutions (MFIs)

KW - k Nearest Neighbours (KNN)

KW - m5

KW - peru

KW - random Forest (RF)

KW - return on Assets (ROA)

KW - wrapper Techniques

UR - https://hdl.handle.net/20.500.12724/12643

UR - http://www.scopus.com/inward/record.url?scp=85100938425&partnerID=8YFLogxK

UR - https://www.mendeley.com/catalogue/5fb62298-c478-3b6a-b2e0-7705c6e9cbce/

U2 - 10.1080/19439342.2021.1884119

DO - 10.1080/19439342.2021.1884119

M3 - Article (Contribution to Journal)

VL - 13

SP - 84

EP - 99

JO - Default journal

JF - Default journal

IS - 1

ER -

Applying wrapper-based variable selection techniques to predict MFIs profitability: evidence from Peru

Resumen

COAR

Categoría OCDE

Temas Repositorio Ulima

Acceder al documento

Otros archivos y enlaces

Huella

Citar esto