土石流発生危険度を機械学習を利用して示した文献のレビューです。
Machine-Learning-Based Prediction Modeling for Debris Flow Occurrence: A Meta-Analysis
要旨
- 84文献を調査（英文のみ）。
- MLに基づく土石流発生予測に関する研究は、計13カ国で実施。
- ベースライン作成として、logistic regression (LR) 、random forest (RF) が一般的。LR(43編)、SVM(24編)、BPNN(21編)、RF(21編)。ただし、直近5年間では、XGBoost、GBDT、CNN、MLP。
- 特徴量エンジニアリング（53編）とモデル比較（46編）が一般的に利用されている。前者は、多重共線性分析(26編)、stepwise(18編)、importance(13編)。
- 解釈手法が活用されることは少ない（16編）。
- ROC‐AUCが最も多く報告された評価指標であった。
- LRはRF、BPNN、SVMに比べて劣っていた。SVMはRFに匹敵。
- データ作成、モデルの構築・評価、予測結果の3つのステップから構成。
- ポリゴン？49編、ポイント36編。ポイントはグリッドセルを利用。
- 特徴量
Category
Description
Topography
Factors related to topography, such as slope, curvature, main channel length, etc.
Morphology
Factors related to the morphology of the surface evaluation unit, such as area, shape coefficient, perimeter, etc.
Geomorphology
Factors related to geomorphic type and evolution, such as landform, hypsometric integra, geomorphic information entropy, etc.
Geology
Factors related to geological structure, geological movement, and geological type, such as active fault density, seismic intensity, lithology, etc.
Meteorology
Factors related to meteorological factors such as rainfall, temperature, snow cover, etc.
Hydrology
Factors related to water flow movement, such as flow accumulation, stream power index, distance to rivers, etc.
Soil
Factors related to soil type, property, and thickness, such as soil texture, soil types, soil depth, etc.
Vegetation
Factors related to vegetation type and state, such as vegetation coverage index, normalized difference vegetation index, forest density, etc.
Fire
Factors related to forest fires, such as fire severity (low, moderate, high), proportion of watersheds burned at high or moderate severity, etc.
Material source
Factors related to loosen accumulation of internal solids, such as collapsed areas, landslide areas, debris reserves, etc.
Human activity
Factors that directly or indirectly characterize human behavior, such as land use, population density, distance to road, etc.
Past debris flow
Factors related to past debris flows in the evaluation unit, such as maximum volume, occurrence frequency, etc.
