TY - JOUR
T1 - The Adjusted SNR and It's Application for Selection Lorenz Function of Income Inequality Analysis
AU - Fajar, Muhammad
AU - Setiawan,
AU - Iriawan, Nur
N1 - Publisher Copyright:
© 2023 The Authors. Published by Elsevier B.V.
PY - 2023
Y1 - 2023
N2 - The formulation of signal-to-noise ratio (SNR) does not take into account the number of parameters used in the model, which can result in capturing intense noise when the number of predictor variables increases. Additionally, there is currently no probabilistic mechanism for comparing the model goodness measure values (in this case, adjusted SNR) of two candidate models used to select the best model. The research objective is to construct an adjusted SNR that accounts for the effect of the number of parameters used in the functional form of a model and establish a Lorenz function selection procedure using the adjusted SNR. The adjusted SNR is constructed by combining SNR and adjusted RR2 (RRaaddjj 2 ), while the Lorenz function selection procedure is formed through a bootstrap mechanism. The proposed methods are applied to household consumption expenditure in Banten Province in 2020, the total income of employment per household in Ghana in 1998, and household income in South Africa in 2015, sourced from BPS-Statistics Banten Province of Indonesia, the Ghana Statistical Service, and Statistics South Africa, respectively. The adjusted SNR developed in this research can be used in empirical practice and addresses the weakness of SNR in not accounting for the number of parameters in the model. Selecting the Lorenz function through bootstrap using the adjusted SNR to measure model goodness-of-fit is a fundamentally new way of analyzing income inequality.
AB - The formulation of signal-to-noise ratio (SNR) does not take into account the number of parameters used in the model, which can result in capturing intense noise when the number of predictor variables increases. Additionally, there is currently no probabilistic mechanism for comparing the model goodness measure values (in this case, adjusted SNR) of two candidate models used to select the best model. The research objective is to construct an adjusted SNR that accounts for the effect of the number of parameters used in the functional form of a model and establish a Lorenz function selection procedure using the adjusted SNR. The adjusted SNR is constructed by combining SNR and adjusted RR2 (RRaaddjj 2 ), while the Lorenz function selection procedure is formed through a bootstrap mechanism. The proposed methods are applied to household consumption expenditure in Banten Province in 2020, the total income of employment per household in Ghana in 1998, and household income in South Africa in 2015, sourced from BPS-Statistics Banten Province of Indonesia, the Ghana Statistical Service, and Statistics South Africa, respectively. The adjusted SNR developed in this research can be used in empirical practice and addresses the weakness of SNR in not accounting for the number of parameters in the model. Selecting the Lorenz function through bootstrap using the adjusted SNR to measure model goodness-of-fit is a fundamentally new way of analyzing income inequality.
KW - Lorenz curve
KW - bootstrap
KW - income
KW - inequality
KW - measure
KW - model selection
KW - signal-to-noise ratio
UR - http://www.scopus.com/inward/record.url?scp=85184348322&partnerID=8YFLogxK
U2 - 10.1016/j.procs.2023.10.497
DO - 10.1016/j.procs.2023.10.497
M3 - Conference article
AN - SCOPUS:85184348322
SN - 1877-0509
VL - 227
SP - 1
EP - 16
JO - Procedia Computer Science
JF - Procedia Computer Science
T2 - 8th International Conference on Computer Science and Computational Intelligence, ICCSCI 2023
Y2 - 2 August 2023 through 3 August 2023
ER -