A Geo-Classification Model for Mapping Mixed Discrete and Continuous
Response Data: An Application to Poverty Mapping

Mark Adjei; Elphas Okango; Richard Puurbalanta; Henry Mwambi; Naiga Babra Charlotte

doi:https://doi.org/10.14445/22315373/IJMTT-V68I1P515

International Journal of Mathematics Trends and Technology

Research Article | Open Access | Download PDF

Volume 68 | Issue 1 | Year 2022 | Article Id. IJMTT-V68I1P515 | DOI : https://doi.org/10.14445/22315373/IJMTT-V68I1P515

A Geo-Classification Model for Mapping Mixed Discrete and Continuous Response Data: An Application to Poverty Mapping

Mark Adjei, Elphas Okango, Richard Puurbalanta, Henry Mwambi, Naiga Babra Charlotte

Citation :

Mark Adjei, Elphas Okango, Richard Puurbalanta, Henry Mwambi, Naiga Babra Charlotte, "A Geo-Classification Model for Mapping Mixed Discrete and Continuous Response Data: An Application to Poverty Mapping," International Journal of Mathematics Trends and Technology (IJMTT), vol. 68, no. 1, pp. 143-157, 2022. Crossref, https://doi.org/10.14445/22315373/IJMTT-V68I1P515

Abstract

Welfare data for monitoring poverty are usually gathered over a wide geographical area, and as such proximal observations are more likely to be affected by common environmental elements and therefore share similar characteristics than distant observations. This is known as spatial dependence. However, poverty analysts have largely ignored this spatial property in welfare data. This work seeks to quantify relationships between poverty-severity and potential covariates while accounting for spatial dependence using a geo-classification model. The source of data for this study, is the seventh round of the Ghana Living Standards Survey (GLSS). We asserted that social and economic characteristics which are bounded in socially constructed spaces affect poverty-generating process. To investigate the interactive association, we use a statistical regime which has the benefit of parsimoniously analyzing all locationspecific circumstances simultaneously, thus yielding a broad view of the processes generating poverty in Ghana. Bayesian estimation was adopted in our model computation. This was due to the hierarchical and highly parameterized nature of our model. Evident from our preliminary results, spatial effect and variation is empirical in the GLSS 7 data and cannot be ignored in the bid to understand poverty and its correlates in the study region. In general, the posterior means and 95% credible intervals show that fixed effect estimates (household size, income level of householder, ecological zone and location/area of residence) and spatial effects significantly influence poverty levels and distribution patterns in Ghana.

Keywords

Household Expenditure, Posterior Densities, Poverty-Severity, Kriging, Spatial Variation.

References

[1] Aboagye-Attah, K. (2019). Socioeconomic correlates of poverty in Ghana using Ghana Living Standards Survey round 6 and 7. page 69.
[2] Alsharkawi, A., Al-Fetyani, M., Dawas, M., Saadeh, H., and Alyaman, M. (2021). Poverty classification using machine learning: The case of Jordan. Sustainability (Switzerland), 13(3):1–16.
[3] Beegle, K., Christiaensen, L., Dabalen, A., and Gaddis, I. (2016). Poverty in a rising Africa. The World Bank.
[4] Ben, R. and Kacem, H. (2012). The Determinants of Poverty by Cohort of Households : Evidence from Rural Tunisia. 36:33–37.
[5] Bogale, A. (2011). Analysis of poverty and its covariates among smallholder farmers in the eastern Hararghe highlands of Ethiopia. Journal of Development and Agricultural Economics, 3(4):157 – 164.
[6] Bukari, C., Essilfie, G., Aning-Agyei, M. A., Otoo, I. C., Kyeremeh, C., Owusu, A. A., Amuquandoh, K. F., and Bukari, K. I. (2021). Impact of COVID-19 on poverty and living standards in Ghana: A micro-perspective. Cogent Economics and Finance, 9(1).
[7] Cakmakyapan, S. and Goktas, A. (2013). a Comparison of Binary Logit and Probit Models With a Simulation Study. Journal of Social and Economic Statistics, 2(1):1–17.
[8] Chambers, R., Salvati, N., and Tzavidis, N. (2016). Semiparametric small area estimation for binary outcomes with application to unemployment estimation for local authorities in the UK. Journal of the Royal Statistical Society. Series A: Statistics in Society.
[9] Cooke, E., Hague, S., and McKay, A. (2016). The Ghana poverty and inequality report: Using the 6th Ghana living standards survey. University of Sussex, pages 1–43.
[10] Cressie, N. A. C. (1993). Statistics for Spatial Data. Journal of the Royal Statistical Society. Series A (Statistics in Society).
[11] Cressie, N. A. C. (1994). Statistics for Spatial Data, Revised Edition. Biometrics.
[12] Darmawan, D. (2019). poverty and shared prosperity 2018: Piecing together the poverty puzzle., volume 53.
[13] De Oliveira, V. (2000). Bayesian prediction of clipped Gaussian random fields. Computational Statistics and Data Analysis, 34(3):299–314.
[14] Dudek, H. and Lisicka, I. (2013). Determinants of poverty – binary logit model with interaction terms approach. Ekonometria, (41):65–77.
[15] Elbers, C., Lanjouw, J. O., and Lanjouw, P. (2003). Micro-level estimation of poverty and inequality. Econometrica, 71(1):355–364.
[16] Ennin, C. C., Nyarko, P. K., Agyeman, A., Mettle, F. O., and Nortey, E. N. N. (2011). Trend analysis of determinants of poverty in Ghana: Logit approach. Research Journal of Mathematics and Statistics, 3(1):20–27.
[17] Gaetan, C. and Guyon, X. (2010). Spatial statistics and modeling, volume 90. Springer.
[18] Gelman, A., Carlin, J. B. B., Stern, H. S. S., and Rubin, D. B. B. (2014). Bayesian Data Analysis, Third Edition (Texts in Statistical Science). Book.
[19] GSS (2019). Ghana Living Standards Survey round 7 (GLSS7), Main Report. Ghana Statistical Service, pages 1–343.
[20] Hobza, T. and Morales, D. (2016). Empirical best prediction under unit-level logit mixed models. Journal of official statistics, 32(3):661.
[21] Matheron, G. (1963). Principles of geostatistics. Economic Geology.
[22] Mdakane, B. P. (2019). Mapping of self-reported health among individuals between the ages of 15 – 49 years in South Africa .
[23] Molina, I., Nandram, B., and Rao, J. N. (2014). Small area estimation of general parameters with application to poverty indicators: A hierarchical bayes approach. Annals of Applied Statistics, 8(2):852–885.
[24] Pleis, J. R. (2018). Mixtures of discrete and continuous variables: Considerations for dimension reduction. pages 1–144.
[25] Puurbalanta, R. (2019). Spatial Cumulative Probit Model: An Application to Poverty Classification and Mapping. International Journal of Statistical Distributions and Applications, 5(1):15.
[26] Puurbalanta, R. (2020). A Clipped Gaussian Geo-Classification model for poverty mapping. Journal of Applied Statistics, 0(0):1–14.
[27] Spiegelhalter, D. J., Best, N. G., Carlin, B. P., and Van Der Linde, A. (2002). Bayesian measures of model complexity and fit. Journal of the Royal Statistical Society. Series B: Statistical Methodology.
[28] Sriyalatha, A. M. K. (2019). Determinants of Poverty Among Households in Monaragala District , Sri Lanka Africa. IRE Journals, 2(7):64–74.
[29] Switzer, P. (1977). Estimation of spatial distributions from point sources with application to air pollution measurement. Technical report No. 9. Technical report.
[30] World Bank (2020). Supporting Countries in Unprecedented Times. Annual Report 2020, pages 1–106