Bayesian spatial models for small area estimation of proportions

Moura, Fernando A. S.; Migon, Hélio S.

doi:10.1191/1471082x02st032oa

Cited by 18 publications

(11 citation statements)

References 14 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…In this paper, we extend these studies by explicitly incorporating spatial correlation structure in our models. Recent research in epidemiology and geography has demonstrated the importance of explicitly modeling such spatial structure in data (Moura and Migon 2002). We model the spatial correlation by imposing complex dependence structures (i.e., spatially correlated prior distributions) on stock-specific pa-rameters.…”

Section: Introductionmentioning

confidence: 99%

Spatial hierarchical Bayesian models for stock-recruitment analysis of pink salmon (Oncorhynchus gorbuscha)

Su¹,

Peterman²,

Haeseker³

2004

Can. J. Fish. Aquat. Sci.

View full text Add to dashboard Cite

To improve the understanding of effects of environmental factors on spawner-to-recruit survival rates of pink salmon (Oncorhynchus gorbuscha), we developed several spatial hierarchical Bayesian models (HBMs). We applied these models to 43 pink salmon stocks in the Northeast Pacific. By using a distance-based, spatially correlated prior distribution for stock-specific parameters, these multistock models explicitly allowed for positive correlation among nearby salmon stocks in their productivities and coefficients of early summer coastal sea surface temperature (SST). To our knowledge, this is the first time that such distance-based, spatial prior probability distributions for parameters have been applied to fisheries problems. We found that the spatial HBMs produce more consistent and precise estimates of effects of SST on productivity than a single-stock approach that estimated parameters for each stock separately. Similar to earlier results using mixed-effects models for the same stocks, we found significant positive effects of SST on survival rates of northern pink salmon stocks, but weaker negative effects of SST on survival rates of southern pink salmon stocks. However, we show a smoother transition in magnitude of effects between these regions.

show abstract

Section: Introductionmentioning

confidence: 99%

Spatial hierarchical Bayesian models for stock-recruitment analysis of pink salmon (Oncorhynchus gorbuscha)

Su¹,

Peterman²,

Haeseker³

2004

Can. J. Fish. Aquat. Sci.

View full text Add to dashboard Cite

show abstract

“…Another approaches have been proposed to handle the issue of non-sampled area. One of the frequently preferred method is the Spatial EBLUP (see [16][17][18][19][20]). This method utilizes spatial dependencies among regions to derive estimates for sampled areas as well as non-sampled areas.…”

Section: Resultsmentioning

confidence: 99%

Improving the quality of disaggregated SDG indicators with cluster information for small area estimates

Zulkarnain¹,

Jayanti²,

Listianingrum³

2020

SJI

View full text Add to dashboard Cite

The increasing needs for more disaggregated data motivates National Statistical Offices (NSOs) to develop efficient methods for producing official statistics without compromising on quality. In Indonesia, regional autonomy requires that Sustainable Development Goals (SDGs) indicators are available up to the district level. However, several surveys such as the Indonesian Demographic and Health Survey produce estimates up to the provincial level only. This generates gaps in support for district level policies. Small area estimation (SAE) techniques are often considered as alternatives for overcoming this issue. SAE enables more reliable estimation of the small areas by utilizing auxiliary information from other sources. However, the standard SAE approach has limitations in estimating non-sampled areas. This paper introduces an approach to estimating the non-sampled area random effect by utilizing cluster information. This model is demonstrated via the estimation of contraception prevalence rates at district levels in North Sumatera province. The results showed that small area estimates considering cluster information (SAE-cluster) produce more precise estimates than the direct method. The SAE-cluster approach revises the direct estimates upward or downward. This approach has important implications for improving the quality of disaggregated SDGs indicators without increasing cost. The paper was prepared under the kind mentorship of Professor James J. Cochran, Associate Dean for Research, Prof. of Statistics and Operations Research, University of Alabama.

show abstract

“…Penambahan unsur spasial berupa spatial random effect bertujuan untuk mengeksplorasi adanya dependensi spasial yang diharapkan dapat meningkatkan efisiensi hasil estimasi dibandingkan hanya dengan menggunakan random effect yang bersifat saling bebas (independen). Penelitian sebelumnya menggunakan model yang sama untuk estimasi proporsi siswa dengan hasil ujian matematika yang masuk dalam grade rendah di Brasil dimana model dengan unsur spasial lebih baik dibandingkan model tanpa unsur spasial (Moura & Migon, 2002). Penelitian selanjutnya, meggunakan model Poisson Log-normal untuk estimasi jumlah bayi dengan berat kurang dari normal (2.500 gram) pada level distrik di Bangladesh dengan model terbaik adalah model dengan unsur random effect independen dan spasial sekaligus (Alam, Hossain, & Sheela, 2019).…”

Section: Pendahuluanunclassified

Penerapan Model Spatial Logit-Normal Pada Small Area Estimation Dengan Metode Hierarchical Bayes

Purwa¹

2020

semnasoffstat

View full text Add to dashboard Cite

Penelitian ini menerapkan model Spatial Logit-normal pada Small Area Estimation (SAE) untuk estimasi proporsi penduduk dengan asupan kalori minimum di bawah 1.400 kkal/kapita/hari pada level kecamatan di Provinsi Bali Tahun 2014 yang merupakan indikator 2.1.2(A) pada tujuan ke-2 SDGs dalam rangka mengukur capaian dan mendukung tercapainya target SDGs pada level lebih tinggi. Terdapat tiga model SAE yang digunakan dengan spesifikasi random effect yang berbeda, yaitu model dengan random effect yang bersifat saling bebas (independen), spatial random effect (iCAR) serta model dengan kedua jenis random effect sekaligus (BYM). Penggunaan unsur spatial random effect diharapkan dapat meningkatkan efisiensi hasil estimasi. Metode estimasi menggunakan pendekatan Hierarchical Bayes (HB) dengan metode Markov Chain Monte Carlo (MCMC) algoritma Gibbs Sampling. Estimasi parameter pada ketiga model menunjukkan hasil yang relatif tidak berbeda dimana hanya ada satu variabel prediktor yang memiliki pengaruh signifikan, yaitu proporsi keluarga pertanian, pada model dengan random effect independen dan model BYM. Sedangkan pada model iCAR tidak ada satu pun variabel prediktor yang berpengaruh signifikan. Berdasarkan nilai Deviance Information Criterion (DIC), model terbaik adalah model BYM. Akan tetapi penambahan unsur spatial random effect bersamaan dengan random effect independen tidak secara signifikan dapat meningkatkan efisiensi hasil estimasi akibat dari minimnya nilai dependensi spasial Moran’s I. Secara visual, pemetaan hasil estimasi dengan model terbaik tidak menunjukkan adanya pola persebaran atau pengelompokan tertentu pada level kecamatan.

show abstract

Bayesian spatial models for small area estimation of proportions

Cited by 18 publications

References 14 publications

Spatial hierarchical Bayesian models for stock-recruitment analysis of pink salmon (Oncorhynchus gorbuscha)

Spatial hierarchical Bayesian models for stock-recruitment analysis of pink salmon (Oncorhynchus gorbuscha)

Improving the quality of disaggregated SDG indicators with cluster information for small area estimates

Penerapan Model Spatial Logit-Normal Pada Small Area Estimation Dengan Metode Hierarchical Bayes

Contact Info

Product

Resources

About