Sponge Species Richness Predictions Using Hybrid Machine Learning Models

Australian marine data was used to model sponge species richness (SSR) with environmental predictors. Researchers applied random forest, generalized linear models, and hybrid geostatistical techniques to generate spatial distribution maps. The study was published in 2017 by authors including Jin Li and affiliated with the Australian Ocean Data Network.

Use Cases

• Predict sponge species richness (SSR) from environmental predictors using random forest models.
• Compare variable selection methods for identifying optimal predictors for species richness models.
• Generate high-accuracy spatial distribution maps of SSR across the study region.
• Analyze the non-linear relationships between sponge richness and seabed features like hard substrate.
• Apply hybrid random forest and geostatistical techniques to model ecological count data.

Strengths

• Study published in a peer-reviewed journal (Environmental Modelling & Software, 2017).
• Methods compared include random forest, GLM, and hybrid geostatistical techniques for model accuracy.

Limitations

• Specific row count, column names, and sample size are not provided in the metadata.
• The underlying raw observation data for sponge richness and environmental variables is not directly available.
• Data is from a 2017 study, which may not reflect current marine conditions.

Provenance

Source

Australian Ocean Data Network.

Collection Method

Modeling study applying machine learning and geostatistics to sponge species and environmental data.

Time Range

null

Freshness

null

Geography

Australian marine regions.