马尔代夫：使用贝叶斯网络为珊瑚礁和气候适应能力提供决策支持（英）2025

World Bank, Digital Development Partnership 2.0 October 2025 Contents 1Executive Summary 6 2Introduction8 2.1Intertwined Ecology and Economy of the Maldives. . . . . . . . . . . . . . 82.2Digital Maldives for Adaptation, Decentralization, and Diversification (DMADD)Project. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .92.3DMADD Project — Component 3. . . . . . . . . . . . . . . . . . . . . . .9 3Structure of the Study10 4Recent Research Innovations and Opportunities114.1Research Innovations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .11 4.1.1Bayesian Networks for Ecosystem Modeling . . . . . . . . . . . . . .114.1.2Reef Monitoring: Remote Sensing and In Situ Observations . . . . .114.1.3Common Findings. . . . . . . . . . . . . . . . . . . . . . . . . . . .124.2Research Opportunities. . . . . . . . . . . . . . . . . . . . . . . . . . . . .124.2.1Dimensionality of Existing Models. . . . . . . . . . . . . . . . . . .124.2.2Causality for Decision Support. . . . . . . . . . . . . . . . . . . . .13 5Theoretical Considerations13 5.1Elements of a Decision Support Tool . . . . . . . . . . . . . . . . . . . . . .135.2The Rationale for Bayesian Networks . . . . . . . . . . . . . . . . . . . . . .155.2.1Bayesian Network Introduction . . . . . . . . . . . . . . . . . . . . .155.2.2Dimensionality and Complexity . . . . . . . . . . . . . . . . . . . . .155.2.3Transparency and Interpretability. . . . . . . . . . . . . . . . . . .165.2.4Probabilistic Approach . . . . . . . . . . . . . . . . . . . . . . . . . .165.2.5Nonparametric Model. . . . . . . . . . . . . . . . . . . . . . . . . .165.2.6Omnidirectional Inference . . . . . . . . . . . . . . . . . . . . . . . .165.2.7Knowledge Integration . . . . . . . . . . . . . . . . . . . . . . . . . .175.2.8Causality. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .175.2.9Utilities. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .185.3The Need for Causality. . . . . . . . . . . . . . . . . . . . . . . . . . . . .18 6Data Collection and Assembly186.1Field Observations. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .19 6.2Secondary Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .206.3Data for Model Development. . . . . . . . . . . . . . . . . . . . . . . . . .20 7Bayesian Network Model Development237.1Machine Learning . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .23 7.1.1Bayesian Network Software. . . . . . . . . . . . . . . . . . . . . . .237.1.2Learning Algorithm. . . . . . . . . . . . . . . . . . . . . . . . . . .237.2Model Limitations. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .247.2.1Observational Inference. . . . . . . . . . . . . . . . . . . . . . . . .247.2.2Complexity. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .247.3Preliminary Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .267.4Effects Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .287.4.1Units of Observation . . . . . . . . . . . . . . . . . . . . . . . . . . .287.4.2Environmental Conditions . . . . . . . . . . . . . . . . . . . . . . . .297.4.3Human Forces. . . . . . . . . . . . . . . . . . . . . . . . . . . . . .327.4.4Marine Life. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .347.5Hierarchical Bayesian Network Model. . . . . . . . . . . . . . . . . . . . .347.5.1Manifest Variables and Latent Factors . . . . . . . . . . . . . . . . .35 8Causal Inference for Policy Analysis: From Prediction to Intervention398.1Causality for Decision Support. . . . . . . . . . . . . . . . . . . . . . . . .39 8.1.1Examples of Policy Questions . . . . . . . . . . . . . . . . . . . . . .398.1.2Examples of Policy Options . . . . . . . . . . . . . . . . . . . . . . .398.2Predictive Models . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .408.3Explanatory/Causal Models . . . . . . . . . . . . . . . . . . . . . . . . . . .408.4Search for a Causal Mechanism . . . . . . . . . . . . . . . . . . . . . . . . .418.5Limitations of Statistical Methods. . . . . . . . . . . . . . . . . . . . . . .418.6Requirements for Causal Inference. . . . . . . . . . . . . . . . . . . . . . .418.7Causal Inference from Observational Data . . . . . . . . . . . . . . . . . . .428.8Assigning Confounders and Non-Confounders. . . . . . . . . . . . . . . . .438.9Causal Driver Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .45 9Decision Support48 9.1Value Judgments. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 489.2Utilities. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .489.3Utility Scales. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .489.4Decision Models. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .499.4.1Decision Model 1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . .499.4.2Decision Model 2 .