Australian Ocean Data Network hosts a dataset from a 2017 study applying random forest, GLM, and hybrid geostatistical methods to predict sponge species richness. The research addressed variable and model selection issues to depict non-linear relationships and generate a high-accuracy spatial distribution of sponge species richness. The dataset likely contains predicted richness values and associated environmental predictors for marine areas.
Use Cases
- Benchmarking predictive models for count data based on the described comparison of random forest, GLM, and hybrid methods.
- Analyzing the relationship between sponge species richness and seabed features based on the revealed association with hard seabed.
- Developing spatial distribution maps for marine species based on the generated high-accuracy predictions.
- Investigating the effect of correlated predictors on model accuracy based on the finding that they may improve random forest performance.
Strengths
- Based on a peer-reviewed study published in Environmental Modelling & Software in 2017.
- Focuses on a high-accuracy spatial prediction method for a key marine ecosystem indicator.
- Addresses specific methodological issues with variable and model selection for ecological count data.
Limitations
- Column-level documentation is absent; field semantics must be inferred after download.
- Row count is unknown, which may limit suitability assessment.
- Last updated 2026-05-05 01:40:53.529535; freshness should be verified.
Provenance
- Source
- Australian Ocean Data Network
- Collection Method
- Model predictions from a study applying statistical and machine learning methods to sponge species richness data.
- Geography
- Australian marine areas, likely.