Figure 2 | Scientific Reports

Figure 2

From: Improving satellite-based PM2.5 estimates in China using Gaussian processes modeling in a Bayesian hierarchical setting

Figure 2

Hierarchical nature and procedures involved in constructing the daily Gaussian processes models. \(i=\mathrm{1..}.,n\) is the location index (\(n\) differs in each daily model); the iterations have indices \(J=1,\mathrm{..}.,5,000\); after the burn-in period, only the iterations after 3,000 (including 3,000) are retained, so \(j=\mathrm{1..}.,2001\). \(\overline{{\omega }_{i}},\overline{{\beta }_{0}},\overline{{\beta }_{1}},\overline{{\sigma }^{2}}\), and \(\bar{\varphi },\overline{{\tau }^{2}}\) are the daily mean values of each parameter. Other symbols have the same definitions as given previously.

Back to article page