EUVIMEDEUVIMED entdecken
Open Access · CC BY-NC-ND 4.0

Regression-Based Identification of Key Determinants in Multivariate Data: Evidence from Global Life Expectancy Indicators

Yogeesh N

Qubahan EcoLead Journal · 21.11.2026 · 5 (4) · 17-28

DOI: 10.48161/qelj.v5n4a61 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Qubahan EcoLead Journal, online erschienen am 21.11.2026. Verfasst von Yogeesh N. Der Volltext ist frei zugänglich (Open Access, CC BY-NC-ND 4.0).

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Multivariate datasets are increasingly used to support evidence-based decision-making, yet their practical value depends on whether researchers can identify which variables most strongly explain an outcome after competing predictors are considered simultaneously. This study develops and applies a regression-based statistical analysis framework for identifying key factors in multivariate data. The empirical case uses an updated global life-expectancy dataset covering 2,864 country-year observations, 179 countries, 9 regions, and 21 variables from 2000 to 2015. Life expectancy was modeled as the dependent variable, while mortality, immunization, disease-burden, economic, demographic, schooling, body-mass, regional, and development-status indicators were used as explanatory variables. The analytical procedure combined descriptive statistics, correlation analysis, least absolute shrinkage and selection operator (LASSO)-assisted variable selection, heteroscedasticity-robust multiple regression, multicollinearity assessment, residual diagnostics, and train-test plus 10-fold cross-validation. The strongest bivariate associations with life expectancy were adult mortality, under-five deaths, infant deaths, schooling, polio immunization, diphtheria immunization, body mass index (BMI), gross domestic product (GDP) per capita, and human immunodeficiency virus (HIV) incidence. The final robust regression model retained mortality, development status, schooling, GDP per capita, BMI, HIV incidence, and regional indicators as key explanatory factors. The validation results showed strong explanatory and predictive performance, with a test coefficient of determination (R²) of 0.982, root mean squared error (RMSE) of 1.232, mean absolute error (MAE) of 0.981, and a 10-fold cross-validated R² of 0.983. The findings demonstrate that regression-based analysis remains a transparent and reliable method for identifying interpretable determinants in multivariate datasets when combined with variable selection, diagnostics, and validation. The study contributes a reproducible empirical framework for key-factor detection in applied statistical research.

Zitieren

APA

N, Y. (2026). Regression-Based Identification of Key Determinants in Multivariate Data: Evidence from Global Life Expectancy Indicators. Qubahan EcoLead Journal, 5(4), 17-28. https://doi.org/10.48161/qelj.v5n4a61

Vancouver

N Y. Regression-Based Identification of Key Determinants in Multivariate Data: Evidence from Global Life Expectancy Indicators. Qubahan EcoLead Journal. 2026;5(4):17-28. doi:10.48161/qelj.v5n4a61

BibTeX
@article{N2026,
  author = {N, Yogeesh},
  title = {Regression-Based Identification of Key Determinants in Multivariate Data: Evidence from Global Life Expectancy Indicators},
  journal = {Qubahan EcoLead Journal},
  year = {2026},
  volume = {5},
  number = {4},
  pages = {17--28},
  doi = {10.48161/qelj.v5n4a61}
}
Kostenlos starten →