Influence of Data Splitting on Performance of Machine Learning Models in Prediction of Shear Strength of Soil

Mathematical Problems in Engineering, Journal Year: 2021, Volume and Issue: 2021, P. 1 - 15

Published: Feb. 5, 2021

The main objective of this study is to evaluate and compare the performance different machine learning (ML) algorithms, namely, Artificial Neural Network (ANN), Extreme Learning Machine (ELM), Boosting Trees (Boosted) considering influence various training testing ratios in predicting soil shear strength, one most critical geotechnical engineering properties civil design construction. For aim, a database 538 samples collected from Long Phu 1 power plant project, Vietnam, was utilized generate datasets for modeling process. Different (i.e., 10/90, 20/80, 30/70, 40/60, 50/50, 60/40, 70/30, 80/20, 90/10) were used divide into assessment models. Popular statistical indicators, such as Root Mean Squared Error (RMSE), Absolute (MAE), Correlation Coefficient (R), employed predictive capability models under ratios. Besides, Monte Carlo simulation simultaneously carried out proposed models, taking account random sampling effect. results showed that although all three ML performed well, ANN accurate statistically stable model after 1000 simulations (Mean R = 0.9348) compared with other Boosted 0.9192) ELM 0.8703). Investigation on greatly affected by training/testing ratios, where 70/30 presented best Concisely, herein an effective manner selecting appropriate predict strength accurately, which would be helpful phases construction projects.

Language: Английский

---

Flood susceptibility modelling using advanced ensemble machine learning models

Geoscience Frontiers, Journal Year: 2020, Volume and Issue: 12(3), P. 101075 - 101075

Published: Oct. 5, 2020

Floods are one of nature's most destructive disasters because the immense damage to land, buildings, and human fatalities. It is difficult forecast areas that vulnerable flash flooding due dynamic complex nature floods. Therefore, earlier identification flood susceptible sites can be performed using advanced machine learning models for managing disasters. In this study, we applied assessed two new hybrid ensemble models, namely Dagging Random Subspace (RS) coupled with Artificial Neural Network (ANN), Forest (RF), Support Vector Machine (SVM) which other three state-of-the-art modelling susceptibility maps at Teesta River basin, northern region Bangladesh. The application these includes twelve influencing factors 413 current former points, were transferred in a GIS environment. information gain ratio, multicollinearity diagnostics tests employed determine association between occurrences influential factors. For validation comparison ability predict statistical appraisal measures such as Freidman, Wilcoxon signed-rank, t-paired Receiver Operating Characteristic Curve (ROC) employed. value Area Under (AUC) ROC was above 0.80 all models. modelling, model performs superior, followed by RF, ANN, SVM, RS, then several benchmark approach solution-oriented outcomes outlined paper will assist state local authorities well policy makers reducing flood-related threats also implementation effective mitigation strategies mitigate future damage.

Language: Английский

---

Performance Evaluation of Machine Learning Methods for Forest Fire Modeling and Prediction

Symmetry, Journal Year: 2020, Volume and Issue: 12(6), P. 1022 - 1022

Published: June 17, 2020

Predicting and mapping fire susceptibility is a top research priority in fire-prone forests worldwide. This study evaluates the abilities of Bayes Network (BN), Naïve (NB), Decision Tree (DT), Multivariate Logistic Regression (MLP) machine learning methods for prediction across Pu Mat National Park, Nghe An Province, Vietnam. The modeling methodology was formulated based on processing information from 57 historical fires set nine spatially explicit explanatory variables, namely elevation, slope degree, aspect, average annual temperate, drought index, river density, land cover, distance roads residential areas. Using area under receiver operating characteristic curve (AUC) seven other performance metrics, models were validated terms their to elucidate general behaviors Park predict future fires. Despite few differences between AUC values, BN model with an value 0.96 dominant over predicting second best DT (AUC = 0.94), followed by NB 0.939), MLR 0.937) models. Our robust analysis demonstrated that these are sufficiently response training validation datasets change. Further, results revealed moderate high levels susceptibilities associated ~19% where human activities numerous. resultant maps provide basis developing more efficient fire-fighting strategies reorganizing policies favor sustainable management forest resources.

Language: Английский

---

Improvement of Best First Decision Trees Using Bagging and Dagging Ensembles for Flood Probability Mapping

Water Resources Management, Journal Year: 2020, Volume and Issue: 34(9), P. 3037 - 3053

Published: June 30, 2020

Language: Английский

---

Groundwater level prediction using machine learning algorithms in a drought-prone area

Neural Computing and Applications, Journal Year: 2022, Volume and Issue: 34(13), P. 10751 - 10773

Published: March 3, 2022

Language: Английский

---

A Study on a Probabilistic Method for Designing Artificial Neural Networks for the Formation of Intelligent Technology Assemblies with High Variability

Electronics, Journal Year: 2023, Volume and Issue: 12(1), P. 215 - 215

Published: Jan. 1, 2023

Currently, ensemble approaches based, among other things, on the use of non-network models are powerful tools for solving data analysis problems in various practical applications. An important problem formation ensembles is ensuring synergy solutions by using properties a variety basic individual solutions; therefore, developing an approach that ensures maintenance diversity preliminary pool relevant development and research. This article devoted to study possibility method probabilistic neural network structures developed authors. In order form networks, influence parameters structure generation quality regression considered. To improve overall solution, flexible adjustment procedure choosing type activation function when filling layers proposed. determine effectiveness this approach, number numerical studies set generated test tasks real datasets were conducted. The forming common solution networks based application evolutionary genetic programming also presents results demonstrate higher efficiency with modified compared selecting best from preformed pool. These carried out several that, particular, describe process ore-thermal melting.

Language: Английский

---

An integrated GIS, MIF, and TOPSIS approach for appraising electric vehicle charging station suitability zones in Mumbai, India

Sustainable Cities and Society, Journal Year: 2023, Volume and Issue: 97, P. 104717 - 104717

Published: June 7, 2023

Language: Английский

---

GIS-based machine learning algorithm for flood susceptibility analysis in the Pagla river basin, Eastern India

Natural Hazards Research, Journal Year: 2023, Volume and Issue: 3(3), P. 420 - 436

Published: May 19, 2023

The unique characteristics of drainage conditions in the Pagla river basin cause flooding and harm socioeconomic environment. main purpose this study is to investigate comparative utility six machine learning algorithms improve flood susceptibility ensemble techniques' capability elucidate underlying patterns floods make a more accurate prediction susceptibilities basin. In present scenario, frequency area becomes high with heavy sudden rainfall, so it essential mitigation measure. At First, spatial database was built 200 locations sixteen influencing factors, its process help Geographic Information System (GIS) environment build up different models applying techniques. It has found zone using learning-based Artificial Neural Network (ANN), Support Vector Machine (SVM), Random Forest (RF), Reduced Error Pruning Tree (REPTree), Logistic Regression (LR), Bagging helping GIS model validation Receiver Operating Characteristic Curve (ROC). Afterward, all gate accuracy zone. calculated under very 8.69%, 14.92%, 14.17%, 12.98%, 14.65%, 13.24% 13.41% for ANN, SVM, RF, REPTree, LR Bagging, respectively. Finally, ROC curve, Standard (SE), Confidence Interval (CI) at 95 per cent were used assess compare performance models. obtained results indicate that are highly accepted Area Under (AUC) between 0.889 (LR) 0.926 (Ensemble). After application, ROC, Ensemble suited highest compared other projecting area. curve AUC values 0.918 0.926, SE (0.023, 034), narrowest CI (95 cent) (0.873–0.962, 0.859–0.993) whereas (the ROC) value (0.914, 0.919), both training datasets. ensembling, result shows susceptible located lower part area, lie 4.46 6.00 result. areas comprise low height belong Murarai I, II, Suti I II C.D. block West Bengal. current will policymakers researcher determine conditioning problems prospects.

Language: Английский

---

Assessing groundwater artificial recharge suitability in the Mi River basin using GIS, RS, and FAHP: a comprehensive analysis with seasonal variations

Applied Water Science, Journal Year: 2025, Volume and Issue: 15(2)

Published: Jan. 29, 2025

Language: Английский

---

Coupling RBF neural network with ensemble learning techniques for landslide susceptibility mapping

CATENA, Journal Year: 2020, Volume and Issue: 195, P. 104805 - 104805

Published: July 25, 2020

Language: Английский

---