# Debris Flow Prediction Based on the Fast Multiple Principal Component Extraction and Optimized Broad Learning

^{1}

^{2}

^{3}

^{4}

^{*}

## Abstract

**:**

## 1. Introduction

## 2. Overview of the Study Area

#### 2.1. Topography and Geomorphology

^{2}and 2400 km

^{2}. Its altitude range is 800~1500 m, located under the Qinling Mountains, with many mountains and gullies, belonging to the middle and low mountain terrain. The mountain’s amount of soil and stone is as high as 1,800,000 km

^{2}accounting for more than 80% of the area, forming a canyon area with significant terrain differences. At the same time, it is located in the middle of the two rivers, with many small river basins and abundant water resources. In summer and autumn, the rainfall is more, and the temperature changes significantly, which is easy to cause loose soil. The region is rich in minerals and the intensity of human activities increases the debris flow safety points.

#### 2.2. Distribution and Law of Debris Flow

## 3. Theories and Method

#### 3.1. FMPCE

**R**. The algorithm form is shown in Equation (4).

**,**$\mathit{F}(k)=\mathit{W}(k)\mathit{A}{\mathit{W}}^{\mathit{T}}(k)\mathit{RW}(k)$.

#### 3.2. Broad Learning

#### 3.2.1. Initial Structure

#### 3.2.2. Updated Structure

#### 3.3. Improved Broad Learning

## 4. Modeling Process and Result Analysis

#### 4.1. Establishment of Debris Flow Probability Prediction Model

Algorithm 1 Debris flow prediction based on the SVDBL algorithm |

Input: Training sample $(\mathit{X},\mathit{Y})$ and actual input $\mathit{Y}$ Output: Debris flow probability $\stackrel{\u2322}{\mathit{Y}}$ |

1 for $\mathrm{i}=0;\mathrm{i}\le \mathrm{n}$ do2 Initialize ${\mathit{W}}_{ei}^{}$ and ${\beta}_{ei}^{}$ 3 Calculate ${\mathit{Z}}_{{}^{i}}={\Phi}_{i}(\mathit{X}{\mathit{W}}_{ei}+{\beta}_{ei})$ 4 Calculate ${\mathit{V}}_{{\mathit{Z}}_{1}}^{\mathit{P}}$ by SVD 5 end6 for $j=0;j\le m$ do7 Initialize ${\mathit{W}}_{{h}_{j}}$ and ${\beta}_{{h}_{j}}$ 8 Calculate ${\mathit{H}}_{j}=\xi \left(\left[{\mathit{Z}}_{1}{\mathit{V}}_{{\mathit{Z}}_{1}}^{\mathit{P}},\dots ,{\mathit{Z}}_{n}{\mathit{V}}_{{\mathit{Z}}_{n}}^{\mathit{P}}\right]{\mathit{W}}_{{h}_{j}}+{\beta}_{{h}_{j}}\right)$ 9 Calculate ${\mathit{V}}_{{\mathit{H}}_{1}}^{\mathit{P}}$ by SVD 10 end11 Let ${\mathit{A}}_{}^{\left\{m,n\right\}}=\left[{\mathit{Z}}_{1}{\mathit{V}}_{{\mathit{Z}}_{1}}^{\mathit{P}},\dots ,{\mathit{Z}}_{n}{\mathit{V}}_{{\mathit{Z}}_{n}}^{\mathit{P}}\left(\right)open="|">{H}_{1}^{},{\mathit{V}}_{{\mathit{H}}_{1}}^{\mathit{P}},,,\dots ,,,{H}_{m}^{},{\mathit{V}}_{{\mathit{H}}_{m}}^{\mathit{P}}\right]$ 12 Calculate ${\left({\mathit{A}}_{}^{\left\{m,n\right\}}\right)}_{}^{+}$ by Equation (8) 13 while the training accuracy does not meet the requirements do 14 Initialize ${\mathit{W}}_{{h}_{m+1}}$ and ${\beta}_{{h}_{m+1}}$ 15 Calculate ${\mathit{H}}_{m+1}=\xi \left(\left[{\mathit{Z}}_{1}{\mathit{V}}_{{\mathit{Z}}_{1}}^{\mathit{P}},\dots ,{\mathit{Z}}_{n}{\mathit{V}}_{{\mathit{Z}}_{n}}^{\mathit{P}}\right]{\mathit{W}}_{{h}_{m+1}}+{\beta}_{{h}_{m+1}}\right)$; Update ${\mathit{A}}_{}^{m+1}$ 16 Calculate ${\mathit{V}}_{{\mathit{H}}_{m+1}}^{\mathit{P}}$ by SVD 17 Update ${\mathit{A}}_{}^{\left\{m+1,n\right\}}$ 18 Calculate ${\left({\mathit{A}}_{\mathit{F}}^{\left\{m+1,n\right\}}\right)}_{}^{+}$ and ${\mathit{W}}_{\mathit{F}}^{\left\{m+1,n\right\}}$ by Equations (22) and (23) 19 $m=m+1$ 20 end21 Calculate ${\mathit{V}}_{\mathit{F}}^{\mathit{P}}$ by SVD 22 Calculate ${\mathit{A}}_{\mathit{F}}^{}={\mathit{A}}_{\mathit{F}}^{\left\{m+1,n\right\}}{\mathit{V}}_{\mathit{F}}^{\mathit{P}}$ 23 Calculate ${\mathit{A}}_{\mathit{F}}^{+}$ and ${\mathit{W}}_{\mathit{F}}^{}={\mathit{A}}_{\mathit{F}}^{+}\mathit{Y}$ by Equation (8) 24 Let $\mathit{W}={\mathit{W}}_{\mathit{F}}^{}$ 25 Calculation of debris flow probability $\stackrel{\u2322}{\mathit{Y}}=\stackrel{\u2322}{\mathit{X}}\mathit{W}$ |

#### 4.2. Determine the Influencing Factors and Data Sources

#### 4.3. Data Preprocessing

- (1)
- Missing value processing

- (2)
- Outlier processing

- (3)
- Normalization

#### 4.4. Evaluating Indicator

^{2}, AUC value and training and test time as indicators.

- (1)
- Root mean square error

- (2)
- Mean absolute percentage error$$MAPE={\displaystyle {\sum}_{i=1}^{n}\left|\frac{y(i)-\overline{y}(i)}{y(i)}\right|}\times \frac{100}{n}$$

- (3)
- coefficient of determination R
^{2}$${R}^{2}=\frac{SSR}{SST}$$

- (4)
- AUC$$AUC=\frac{{\displaystyle {\sum}_{}^{}rank{k}_{i}-\frac{M(M+1)}{2}}}{M\times N}$$

#### 4.5. Simulation Verification and the Result Analysis

^{2}in the performance evaluation of the four models. The RMSE of SVDBL is 0.1309, 0.0818 and 0.0144, smaller than that of GD-BP, SVM and BL, respectively. MAPE of SVDBL is 0.0314, 0.0256 and 0.0190 smaller than that of GD-BP, SVM and BL, respectively. Regarding goodness of fit R2, SVDBL is 0.0777, 0.1020 and 0.0309 higher than the above three models, respectively. This shows that the proposed SVDBL model has the best performance.

^{2}. The results are shown in Figure 12 and Figure 13.

^{2}of the overall model first shows a decreasing trend and a slow and stable trend. The R

^{2}of GD-BP is basically below 0.75. The support vector machine SVM is superior to GD-BP, and the R

^{2}is slightly higher than 0.75. The R

^{2}of BL is between 0.75 and 0.8, and the performance is somewhat outstanding for the support vector machine SVM. The SVDBL is better than other models, and the R

^{2}is maintained above 0.80, indicating that the model works best.

## 5. Conclusions

## Author Contributions

## Funding

## Data Availability Statement

## Acknowledgments

## Conflicts of Interest

## References

- Reid, M.E.; Coe, J.A.; Brien, D.L. Forecasting Inundation from Debris Flows That Grow Volumetrically during Travel, with Application to the Oregon Coast Range, USA. Geomorphology
**2016**, 273, 396–411. [Google Scholar] [CrossRef] [Green Version] - Li, L.; Cheng, S.; Wen, Z. Landslide Prediction Based on Improved Principal Component Analysis and Mixed Kernel Function Least Squares Support Vector Regression Model. J. Mt. Sci.
**2021**, 18, 2130–2142. [Google Scholar] [CrossRef] - Kim, H.-S.; Chung, C.-K.; Kim, S.-R.; Kim, K.-S. A GIS-Based Framework for Real-Time Debris-Flow Hazard Assessment for Expressways in Korea. Int. J. Disaster Risk Sci.
**2016**, 7, 293–311. [Google Scholar] [CrossRef] [Green Version] - Zahiri, A.; Najafzadeh, M. Optimized Expressions to Evaluate the Flow Discharge in Main Channels and Floodplains Using Evolutionary Computing and Model Classification. Int. J. River Basin Manag.
**2018**, 16, 123–132. [Google Scholar] [CrossRef] - Gong, X.-L.; Chen, X.-Q.; Chen, K.-T.; Zhao, W.-Y.; Chen, J.-G. Engineering Planning Method and Control Modes for Debris Flow Disasters in Scenic Areas. Front. Earth Sci.
**2021**, 9, 712403. [Google Scholar] [CrossRef] - Zhu, X.H.; Zhuang, J.Q.; Liu, T.M.; Tong, X.; Ma, P.H. Developmental characteristics and hazard analysis of debris flow in Jiaochang Gully of Huxian County, Xi’an City. J. Nat. Disasters
**2016**, 25, 27–33. [Google Scholar] [CrossRef] - Tiwari, B.; Ajmera, B.; Gonzalez, A.; Sonbol, H. Impact of Wildfire on Triggering Mudslides—A Case Study of 2018 Montecito Debris Flows. In Proceedings of the Geo-Congress 2020: Engineering, Monitoring, and Management of Geotechnical Infrastructure, Minneapolis, MN, USA, 25–28 February 2020; American Society of Civil Engineers (ASCE): Reston, VA, USA, 2020; pp. 40–49. [Google Scholar]
- Zhang, H.; Liu, X.; Cai, E.; Huang, G.; Ding, C. Integration of Dynamic Rainfall Data with Environmental Factors to Forecast Debris Flow Using an Improved GMDH Model. Comput. Geosci.
**2013**, 56, 23–31. [Google Scholar] [CrossRef] - Pudasaini, S.P.; Fischer, J.-T. A Mechanical Model for Phase Separation in Debris Flow. Int. J. Multiph. Flow
**2020**, 129, 103292. [Google Scholar] [CrossRef] [Green Version] - Xu, W.; Yu, W.; Zhang, G. Prediction Method of Debris Flow by Logistic Model with Two Types of Rainfall: A Case Study in the Sichuan, China. Nat. Hazards
**2012**, 62, 733–744. [Google Scholar] [CrossRef] - Kim, S.D.; Jeung, S.J.; Kim, B.S. Prediction of Debris Flows in the Korean Oship River Based on Climate Change Scenarios. Geomat. Nat. Hazards Risk
**2018**, 9, 703–719. [Google Scholar] [CrossRef] - Lombardo, L.; Bachofer, F.; Cama, M.; Märker, M.; Rotigliano, E. Exploiting Maximum Entropy Method and ASTER Data for Assessing Debris Flow and Debris Slide Susceptibility for the Giampilieri Catchment (North-Eastern Sicily, Italy). Earth Surf. Process. Landf.
**2016**, 41, 1776–1789. [Google Scholar] [CrossRef] [Green Version] - Najafzadeh, M. Neuro-Fuzzy GMDH Systems Based Evolutionary Algorithms to Predict Scour Pile Groups in Clear Water Conditions. Ocean. Eng.
**2015**, 99, 85–94. [Google Scholar] [CrossRef] - Di, B.; Zhang, H.; Liu, Y.; Li, J.; Chen, N.; Stamatopoulos, C.A.; Luo, Y.; Zhan, Y. Assessing Susceptibility of Debris Flow in Southwest China Using Gradient Boosting Machine. Sci. Rep.
**2019**, 9, 12532. [Google Scholar] [CrossRef] [Green Version] - Pandey, V.K.; Sharma, K.K.; Pourghasemi, H.R.; Bandooni, S.K. Sedimentological Characteristics and Application of Machine Learning Techniques for Landslide Susceptibility Modelling along the Highway Corridor Nahan to Rajgarh (Himachal Pradesh), India. CATENA
**2019**, 182, 104150. [Google Scholar] [CrossRef] - Zhang, S.H.; Wu, G. Debris Flow Susceptibility and Its Reliability Based on Random Forest and GIS. Sci. J. Earth Sci.
**2019**, 44, 3115–3134. [Google Scholar] [CrossRef] - Nikolopoulos, E.I.; Destro, E.; Bhuiyan, M.A.E.; Borga, M.; Anagnostou, E.N. Evaluation of Predictive Models for Post-Fire Debris Flow Occurrence in the Western United States. Nat. Hazards Earth Syst. Sci.
**2018**, 18, 2331–2343. [Google Scholar] [CrossRef] [Green Version] - Liu, D.L.; Zhang, S.J.; Yang, H.J. Design and implementation of debris-flow forecast system based on water-soil coupling mechanism. Comput. Eng. Des.
**2014**, 35, 3667–3673. [Google Scholar] [CrossRef] - Geng, J.D.; Chen, W.Y.; Zhao, Y.H. A Precise Prediction Model on Debris Flows Caused by Runoff Mechanism Based on Channel Width and Particle Size. China Earthq. Eng. J.
**2021**, 43, 983–988. [Google Scholar] - Banihabib, M.E.; Tanhapour, M. An Empirical Equation to Determine the Threshold for Rainfall-Induced Landslides Developing to Debris Flows. Landslides
**2020**, 17, 2055–2065. [Google Scholar] [CrossRef] - Wei, Z.L.; Sun, H.Y.; Pan, P.; Shang, Y.Q. Forecast for average velocity of debris flow based on MLS method. J. Nat. Disasters
**2016**, 25, 56–61. [Google Scholar] [CrossRef] - Imaizumi, F.; Masui, T.; Yokota, Y.; Tsunetaka, H.; Hayakawa, Y.S.; Hotta, N. Initiation and Runout Characteristics of Debris Flow Surges in Ohya Landslide Scar, Japan. Geomorphology
**2019**, 339, 58–69. [Google Scholar] [CrossRef] - Vergara Dal Pont, I.P.; Santibañez Ossa, F.A.; Araneo, D.; Ferrando Acuña, F.J.; Moreiras, S.M. Determination of Probabilities for the Generation of High-Discharge Flows in the Middle Basin of Elqui River, Chile. Nat. Hazards
**2018**, 93, 531–546. [Google Scholar] [CrossRef] - Giannecchini, R.; Galanti, Y.; D’Amato Avanzi, G.; Barsanti, M. Probabilistic Rainfall Thresholds for Triggering Debris Flows in a Human-Modified Landscape. Geomorphology
**2016**, 257, 94–107. [Google Scholar] [CrossRef] - Meng, F.Q.; Li, G.J.; Wang, Q.B.; Qing, S.W.; Zhao, H.Q.; Jin, X. Research on early warning of debris flow based on efficacy coefficient method. Rock Soil Mech.
**2012**, 33, 835–840. [Google Scholar] [CrossRef] - Zhang, S.; Yang, H.; Wei, F.; Jiang, Y.; Liu, D. A Model of Debris Flow Forecast Based on the Water-Soil Coupling Mechanism. J. Earth Sci.
**2014**, 25, 757–763. [Google Scholar] [CrossRef] - Liang, W.Z.; Zhao, G.Y. Comprehensive prediction of mine debris flow’risk based on weight-varying cloud model. Chin. J. Geol. Hazard Control.
**2017**, 28, 82–88. [Google Scholar] [CrossRef] - Qiao, S.; Qin, S.; Chen, J.; Hu, X.; Ma, Z. The Application of a Three-Dimensional Deterministic Model in the Study of Debris Flow Prediction Based on the Rainfall-Unstable Soil Coupling Mechanism. Processes
**2019**, 7, 99. [Google Scholar] [CrossRef] [Green Version] - Banihabib, M.E.; Tanhapour, M.; Roozbahani, A. Bayesian Networks Model for Identification of the Effective Variables in the Forecasting of Debris Flows Occurrence. Environ. Earth Sci.
**2020**, 79, 179. [Google Scholar] [CrossRef] - Wu, M.-H.; Wang, J.P.; Chen, I.-C. Optimization Approach for Determining Rainfall Duration-Intensity Thresholds for Debris Flow Forecasting. Bull. Eng. Geol. Environ.
**2019**, 78, 2495–2501. [Google Scholar] [CrossRef] - Kern, A.N.; Addison, P.; Oommen, T.; Salazar, S.E.; Coffman, R.A. Machine Learning Based Predictive Modeling of Debris Flow Probability Following Wildfire in the Intermountain Western United States. Math. Geosci.
**2017**, 49, 717–735. [Google Scholar] [CrossRef] - Xiao, L.; Zhang, Y.; Ge, T.; Wang, C.; Wei, M. Analysis, Assessment and Early Warning of Mudflow Disasters along the Shigatse Section of the China–Nepal Highway. Open Geosci.
**2020**, 12, 44–58. [Google Scholar] [CrossRef] - Su, T.-J.; Pan, T.-S.; Chang, Y.-L.; Lin, S.-S.; Hao, M.-J. A Hybrid Fuzzy and K-Nearest Neighbor Approach for Debris Flow Disaster Prevention. IEEE Access
**2022**, 10, 21787–21797. [Google Scholar] [CrossRef] - Hirschberg, J.; Badoux, A.; McArdell, B.W.; Leonarduzzi, E.; Molnar, P. Evaluating Methods for Debris-Flow Prediction Based on Rainfall in an Alpine Catchment. Nat. Hazards Earth Syst. Sci.
**2021**, 21, 2773–2789. [Google Scholar] [CrossRef] - Wang, X.; Wang, C.; Zhang, C. Early Warning of Debris Flow Using Optimized Self-Organizing Feature Mapping Network. Water Supply
**2020**, 20, 2455–2470. [Google Scholar] [CrossRef] - Cao, C.; Song, S.; Chen, J.; Zheng, L.; Kong, Y. An Approach to Predict Debris Flow Average Velocity. Water
**2017**, 9, 205. [Google Scholar] [CrossRef] [Green Version] - Zhang, Y.; Wu, K.L.; Deng, X.Q.; Wang, W. Prediction model of Jiangjiagou debris flow average velocity based on relevance vector machine. J. Nat. Disasters
**2019**, 28, 146–153. [Google Scholar] [CrossRef] - Lee, D.-H.; Cheon, E.; Lim, H.-H.; Choi, S.-K.; Kim, Y.-T.; Lee, S.-R. An Artificial Neural Network Model to Predict Debris-Flow Volumes Caused by Extreme Rainfall in the Central Region of South Korea. Eng. Geol.
**2021**, 281, 105979. [Google Scholar] [CrossRef] - Mei, Y.; Hong, F.; Jia, Z.; Kang, Z. Debris Flow Forecasting of Northwest of Yunnan Province Based on LR, SVM, and RF Statistical Models. In Proceedings of the 2018 26th International Conference on Geoinformatics, Kunming, China, 28–30 June 2018; pp. 1–7. [Google Scholar]
- Hu, Y.H.; Yuan, L.; Ma, D.T.; Li, M. Research Progress on Debris Flow Infrasound Warning. Adv. Earth Sci.
**2018**, 33, 606–613. [Google Scholar] - Li, Y.; Chen, W.; Rezaie, F.; Rahmati, O.; Davoudi Moghaddam, D.; Tiefenbacher, J.; Panahi, M.; Lee, M.-J.; Kulakowski, D.; Tien Bui, D.; et al. Debris Flows Modeling Using Geo-Environmental Factors: Developing Hybridized Deep-Learning Algorithms. Geocarto Int.
**2022**, 37, 5150–5173. [Google Scholar] [CrossRef] - Min, Y.; Hu, J.; Li, C.; Xu, Y.J.; Li, H.; Li, X.; Wan, S.; Li, L.; Yang, S.Y. Prediction Model about Landslide and Debris Flow in Yunnan Province. J. Catastrophol.
**2013**, 28, 216–220. [Google Scholar] - Kean, J.W.; Staley, D.M. Forecasting the Frequency and Magnitude of Postfire Debris Flows Across Southern California. Earth Future
**2021**, 9, e2020EF001735. [Google Scholar] [CrossRef] - Kong, L.Z.; Gao, Y.B.; Li, H.; Zhang, H.P. A Fast Algorithm That Extracts Multiple Principle Components in Parallel. Acta Autom. Sin.
**2017**, 43, 835–842. [Google Scholar] [CrossRef] - Chen, C.L.P.; Liu, Z. Broad Learning System: An Effective and Efficient Incremental Learning System Without the Need for Deep Architecture. IEEE Trans. Neural Netw. Learn. Syst.
**2018**, 29, 10–24. [Google Scholar] [CrossRef] [PubMed] - Xu, A.-B.; Xie, D. Low-Rank Approximation Pursuit for Matrix Completion. Mech. Syst. Signal Processing
**2017**, 95, 77–89. [Google Scholar] [CrossRef] - Marra, F.; Nikolopoulos, E.I.; Creutin, J.D.; Borga, M. Space–Time Organization of Debris Flows-Triggering Rainfall and Its Effect on the Identification of the Rainfall Threshold Relationship. J. Hydrol.
**2016**, 541, 246–255. [Google Scholar] [CrossRef] - Mostbauer, K.; Kaitna, R.; Prenner, D.; Hrachowitz, M. The Temporally Varying Roles of Rainfall, Snowmelt and Soil Moisture for Debris Flow Initiation in a Snow-Dominated System. Hydrol. Earth Syst. Sci.
**2018**, 22, 3493–3513. [Google Scholar] [CrossRef] [Green Version] - Tayyebi, S.M.; Pastor, M.; Stickle, M.M. Two-Phase SPH Numerical Study of Pore-Water Pressure Effect on Debris Flows Mobility: Yu Tung Debris Flow. Comput. Geotech.
**2021**, 132, 103973. [Google Scholar] [CrossRef] - Morino, C.; Conway, S.J.; Balme, M.R.; Hillier, J.; Jordan, C.; Sæmundsson, Þ.; Argles, T. Debris-Flow Release Processes Investigated through the Analysis of Multi-Temporal LiDAR Datasets in North-Western Iceland: Debris-Flow Processes Investigated by Multi-Temporal LiDAR Datasets. Earth Surf. Process. Landf.
**2019**, 44, 144–159. [Google Scholar] [CrossRef]

Number | Influence Factors |
---|---|

1 | Rainfall |

2 | Soil moisture content |

3 | Pore water pressure |

4 | Watershed development degree |

5 | Watershed integrity coefficient |

6 | Vegetation coverage |

7 | Slope gradient |

8 | Lithologic |

9 | Watershed area |

10 | Relative height difference |

11 | Erosion and deposition amplitude |

12 | Gully bed gradient |

13 | Recharge section length ratio |

14 | Loose material reserves along the gully |

15 | Adverse geologic phenomena |

16 | New structure influence |

Influence Factors | Representation Significance | The Probability of Debris Flow/% | ||||
---|---|---|---|---|---|---|

<20 | 20~40 | 40~60 | 60~80 | >80 | ||

Rainfall/mm | Water source conditions required for debris flow formation, expressed by comprehensive daily rainfall | <20 | 20~50 | 50~75 | 75~100 | >100 |

Slope gradient/% | The potential energy conditions required for the collection of loose solid matter to quantify the ratio of 25° to 45° area to the basin area | <0.05 | 0.05~0.1 | 0.1~0.15 | 0.15~0.2 | >0.2 |

Gully bed gradient/‰ | Potential Energy Conditions for Debris Flow Formation | <50 | 50~100 | 100~200 | 200~400 | >400 |

Relative height difference/m | Terrain Factors of Debris Flow Formation | <20 | 20~40 | 40~60 | 60~80 | >80 |

Soil moisture content/% | The amount of water in the soil | <8 | 8~16 | 16~24 | 24~32 | >32 |

Pore water pressure/KPa | The pressure of groundwater in soil or rock, acting between particles or pores. | <10 | 10~25 | 25~40 | 40~60 | >60 |

Number | Impact Factor | ||
---|---|---|---|

Rainfall | Soil Moisture Content | Pore Water Pressure | |

1 | 18.32 | 6 | 0.20 |

2 | 23.28 | 13 | 0.13 |

3 | 17.62 | 20 | 0.48 |

4 | 38.74 | 27 | −0.26 |

… | … | … | … |

2000 | 138.62 | 33 | 0.74 |

Before Data Preprocessing | After Data Preprocessing | ||||
---|---|---|---|---|---|

Incremental nodes/Pcs | Accuracy/% | Training time/s | Incremental nodes/Pcs | Accuracy/% | Training time/s |

7000 | 80 | 5.8124 | 2000 | 93 | 1.2216 |

Model | Evaluating Indicators | ||
---|---|---|---|

RMSE | MAPE | R^{2} | |

GD-BP | 0.3706 | 0.0526 | 0.7826 |

SVM | 0.3215 | 0.0468 | 0.7583 |

BL | 0.2541 | 0.0022 | 0.8294 |

SVDBL | 0.2397 | 0.0212 | 0.8603 |

Model | GD-BP | SVM | BL | SVDBL |
---|---|---|---|---|

Average accuracy/% | 85.5 | 90 | 92.5 | 93 |

Average training time/s | 23.2906 | 18.1740 | 2.8899 | 1.2216 |

Warning Level | Possibility of Disaster | Color Identification | Explanation | Probability |
---|---|---|---|---|

I | Very low | Green | Only send information to decision-makers | <20% |

II | Lower | Blue | Push information to decision-makers and related technical personnel | 20%~40% |

III | Medium | Huang | Recommend preventive measures | 40%~60% |

IV | Higher | Orange | Take preventive measures | 60%~80% |

V | Extremely high | Red | Organizing a public emergency response | >80% |

Model | Average AUC Value/% | Average Training Time/s |
---|---|---|

GD-BP | 80.41% | 0.0817 |

SVM | 85.52% | 0.0049 |

BL | 90.37% | 0.0022 |

SVDBL | 93.16% | 0.0010 |

Model | Results of before Training Sample Expansion | Results of after Training Sample Expansion |
---|---|---|

Training Time/s | Training Time/s | |

GD-BP | 23.2906 | 31.9558 |

SVM | 18.1740 | 22.9869 |

BL | 2.8899 | 4.0573 |

SVDBL | 1.2216 | 1.2299 |

**Table 10.**Correspondence between the influence factor and the curve in Figure 7.

Artificially Selected Impact Factors | The Corresponding Curve in Figure 7 |
---|---|

1,2,3,7,12,14 | b |

1,2,3,4,5,12 | c |

1,2,7,8,10,12 | d |

1,3,6,7,10,14 | e |

Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations. |

© 2022 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https://creativecommons.org/licenses/by/4.0/).

## Share and Cite

**MDPI and ACS Style**

Xu, G.; Yan, X.-E.; Cao, N.; Ma, J.; Xie, G.; Li, L.
Debris Flow Prediction Based on the Fast Multiple Principal Component Extraction and Optimized Broad Learning. *Water* **2022**, *14*, 3374.
https://doi.org/10.3390/w14213374

**AMA Style**

Xu G, Yan X-E, Cao N, Ma J, Xie G, Li L.
Debris Flow Prediction Based on the Fast Multiple Principal Component Extraction and Optimized Broad Learning. *Water*. 2022; 14(21):3374.
https://doi.org/10.3390/w14213374

**Chicago/Turabian Style**

Xu, Genqi, Xin-E Yan, Ning Cao, Jing Ma, Guokun Xie, and Lu Li.
2022. "Debris Flow Prediction Based on the Fast Multiple Principal Component Extraction and Optimized Broad Learning" *Water* 14, no. 21: 3374.
https://doi.org/10.3390/w14213374