Comparative Feature Selection Methods for Flood Susceptibility Mapping in Khuzestan, Iran

A 5.5 KB Excel table compares nine feature selection methods applied to flood susceptibility mapping. The dataset was created by Mohammad Kazemi and last updated on 2026-04-29. It is based on 19 initial environmental factors and 1,000 sample points from flood-prone Khuzestan Province, Iran.

Use Cases

• Benchmarking feature selection algorithm performance based on the nine methods listed in the description
• Training metaheuristic-optimized models for flood prediction based on the described LSTM framework
• Identifying critical environmental predictors for flooding based on the consensus selection of NDVI and daily minimum temperature
• Replicating the integrated feature selection and optimization framework for other geospatial hazard mapping tasks

Strengths

• Compares nine distinct feature selection methods, including Boruta, SHAP, and Deep Feature Importance
• Based on 1,000 sample points (500 flood and 500 non-flood) from a defined study area
• Documents a specific performance outcome, with the LSTM-WOA model achieving an F1-Score of 0.88

Limitations

• Row count and column-level documentation are unknown, which may limit suitability assessment
• The dataset is very small (5.5 KB), indicating limited scope, likely a summary table rather than raw data
• Geographic coverage is limited to Khuzestan Province, Iran, which may affect generalizability

Provenance

Source

figshare, author Mohammad Kazemi

Collection Method

Features were sourced from Google Earth Engine, with model training on sampled points generalized to the study area.

Freshness

Last updated 2026-04-29 17:42:20; freshness should be verified

Geography

Khuzestan Province, Iran