Interpretable Predictive Modelling of Basalt Fiber Reinforced Concrete Splitting Tensile Strength Using Ensemble Machine Learning Methods and SHAP Approach

Materials, Journal Year: 2023, Volume and Issue: 16(13), P. 4578 - 4578

Published: June 25, 2023

Basalt fibers are a type of reinforcing fiber that can be added to concrete improve its strength, durability, resistance cracking, and overall performance. The addition basalt with high tensile strength has particularly favorable impact on the splitting concrete. current study presents data set experimental results tests curated from literature. Some best-performing ensemble learning techniques such as Extreme Gradient Boosting (XGBoost), Light Machine (LightGBM), Random Forest, Categorical (CatBoost) have been applied prediction reinforced fibers. State-of-the-art performance metrics root mean squared error, absolute error coefficient determination used for measuring accuracy prediction. each input feature model visualized using Shapley Additive Explanations (SHAP) algorithm individual conditional expectation (ICE) plots. A greater than 0.9 could achieved by XGBoost in strength.

Language: Английский

---

How accurate are the machine learning models in improving monthly rainfall prediction in hyper arid environment?

Journal of Hydrology, Journal Year: 2024, Volume and Issue: 633, P. 131040 - 131040

Published: March 11, 2024

Language: Английский

---

Data-driven novel deep learning applications for the prediction of rainfall using meteorological data

Frontiers in Environmental Science, Journal Year: 2024, Volume and Issue: 12

Published: Aug. 16, 2024

Rainfall plays an important role in maintaining the water cycle by replenishing aquifers, lakes, and rivers, supporting aquatic life, sustaining terrestrial ecosystems. Accurate prediction is crucial given intricate interplay of atmospheric oceanic phenomena, especially amidst contemporary challenges. In this study, to predict rainfall, 12,852 data points from open-source global weather for three cities Indonesia were utilized, incorporating input variables such as maximum temperature (°C), minimum wind speed (m/s), relative humidity (%), solar radiation (MJ/m 2 ). Three novel robust Deep Learning models used: Recurrent Neural Network (DRNN), Gated Unit (DGRU), Long Short-Term Memory (DLSTM). Evaluation results, including statistical metrics like Root-Mean-Square Errors Correction Coefficient (R ), revealed that model outperformed DRNN with values 0.1289 0.9995, respectively. DLSTM networks offer several advantages rainfall prediction, particularly sequential time series excelling handling long-term dependencies capturing patterns over extended periods. Equipped memory cell architecture forget gates, effectively retain retrieve relevant information. Furthermore, enable parallelization, enhancing computational efficiency, flexibility design regularization techniques improved generalization performance. Additionally, results indicate parameters exhibit indirect influence on while temperature, speed, have a direct relationship rainfall.

Language: Английский

---

Long-term Rainfall Forecasting in Arid Climates Using Artificial Intelligence and Statistical Recurrent Models

Journal of Engineering Research, Journal Year: 2024, Volume and Issue: unknown

Published: March 1, 2024

Rainfall is a major component of the hydrologic cycle and thus requires comprehensive understanding its dynamics variability. This study aims to develop test applicability recurrent models for forecasting rainfall in extremely arid regions on monthly time scale. Specifically, Neural Auto-regressive Networks (NARs) Integrated Moving Average (ARIMA) were utilized modeling dataset from Kuwait City 1958 2018. The site possesses extreme conditions with long-term average annual less than 120 mm. harsh condition imposes challenges efforts. results showed that NAR model was more efficient over period. A notable bias encountered within abnormal wet seasons. efforts presented this found be reasonable, they qualify making objective forecasts area other similar climatic zones. overall Nash–Sutcliffe (NS) coefficient 0.206 model, showing an even better performance medium-to-low intensity months (<30 mm per month). With outcome study, operational framework water managers hyper-arid zones aid developing resilient management plans cope adverse impacts climate change.

Language: Английский

---

The State of Art in Machine Learning Applications in Civil Engineering

Studies in systems, decision and control, Journal Year: 2023, Volume and Issue: unknown, P. 147 - 177

Published: Jan. 1, 2023

Language: Английский

---

Revolutionizing Saudi Arabia's Agriculture: The IoT Transformation of Water Management

Journal of Advanced Research in Applied Sciences and Engineering Technology, Journal Year: 2023, Volume and Issue: 36(1), P. 217 - 240

Published: Dec. 24, 2023

Saudi Arabia's agriculture heavily depends on effective water management, given its limited freshwater resources and arid climate. Real-time monitoring of soil moisture levels, weather conditions, crop watering needs, facilitated by IoT integration, plays a crucial role in conserving minimizing waste. The resultant improvements yields quality are essential for the long-term success country. This study employs Technique Order Preference Similarity to Ideal Solution (TOPSIS) method investigate transformative potential Internet Things (IoT) enhancing management practices sector. research begins highlighting significance agriculture, emphasizing proportion land Arabia allocated agricultural purposes. problem statement underscores pressing challenges encompassing issues such as scarcity, inefficient irrigation methods, need real-time data inform decision-making. To address these challenges, proposes an IoT-based Agricultural Water Management System (IoT-AWMS) that leverages sensors, analytics, machine learning algorithms. system is designed optimize utilization agriculture. Simulations conducted within demonstrate significant enhancement usage efficiency, resulting reduced wastage increased yields. In conclusion, this critical importance proposed Arabia. It positioned valuable tool mitigating scarcity promoting environmentally sustainable

Language: Английский

---

Hybrid Drought Forecasting Framework for Water‐Scarce Regions Based on Support Vector Machine and Precipitation Index

Hydrological Processes, Journal Year: 2024, Volume and Issue: 38(12)

Published: Dec. 1, 2024

ABSTRACT Drought is a natural event that slowly deteriorates water reserves. This study aims to develop machine learning–based computational framework for monitoring drought status in water‐scarce regions. The proposed integrates the precipitation index (PI) with support vector models forecast occurrences based on an autoregressive modelling scheme. Due suitability of PI analysis arid climates, developed hybrid model appropriate regions limited rainfall. used historical dataset from 1958 2020 at Kuwait International Airport, City. area characterised by scarce rainfall and vulnerable severe shortages owing resources. Initially, time‐series datasets were examined stationarity validate utility model. autocorrelation function test was significantly associated time series 12‐ 24‐month drought‐monitoring scales. Predictive forecasting constructed predict up 3 months advance. Statistical evaluation metrics assess performance results showed strong association between observed predicted events, coefficients determination ( R 2 ) ranging 0.865 0.925 provide managers efficient reliable tools assist preparing management plans. provides guidance improving resource resilience under shortage scenarios other climatic applying suitable indices conjunction robust data‐driven models. baseline policymakers worldwide establish sustainable conservation strategies crucial insights disaster preparation.

Language: Английский

---

An instructional emperor pigeon optimization (IEPO) based DeepEnrollNet for university student enrolment prediction and retention recommendation

Scientific Reports, Journal Year: 2024, Volume and Issue: 14(1)

Published: Dec. 28, 2024

Academic institutions face increasing challenges in predicting student enrollment and managing retention. A comprehensive strategy is required to track progress, predict future course demand, prevent churn across various disciplines. Institutions need an effective method while addressing potential churn. The existing approaches are often inadequate handling both numerical textual data, limiting the ability provide personalized retention strategies. We propose innovative framework that combines deep learning with recommender systems for prediction prevention. integrates advanced preprocessing techniques numeric data. Feature extraction performed statistical measures text like GloVe embeddings, Latent Dirichlet Allocation (LDA) topic modeling, SentiWordNet sentiment analysis. weighted feature fusion approach these features, optimal features selected using Pythagorean fuzzy AHP a Hybrid Optimization approach, specifically Instructional Emperor Pigeon (IEPO). DeepEnrollNet model, hybrid CNN-GRU-Attention QCNN architecture, used prediction, Deep Q-Networks (DQN) applied generate actionable recommendations. This methodology improves predictive accuracy enrolment provides tailored strategies enhance by data unified framework. has minimum MSE of 0.218978, MSRE 0.216445, NMSE 0.232453, RMSE 0.23213, MAPE 0.218754.

Language: Английский

---

A Framework for Agriculture Plant Disease Prediction using Deep Learning Classifier

International Journal of Advanced Computer Science and Applications, Journal Year: 2023, Volume and Issue: 14(8)

Published: Jan. 1, 2023

The agricultural industry in Saudi Arabia suffers from the effects of vegetable diseases Central Province. primary causes death documented this analysis were 32 fungal diseases, two viral physiological and one parasitic disease. Because early diagnosis plant may boost productivity quality operations, tomatoes, Pepper Onion selected for experiment. goal is to fine-tune hyperparameters common Machine Learning classifiers Deep architectures order make precise diagnoses diseases. first stage makes use image processing methods using ml classifiers; input picture median filtered, contrast increased, background removed HSV color space segmentation. After shape, texture, features have been extracted feature descriptors, hyperparameter-tuned machine learning (ML) such as k-nearest neighbor, logistic regression, support vector machine, random forest are used determine an outcome. Finally, proposed Plant Disease Detection System (DLPDS) Tuned ML models. In second stage, potential Convolutional Neural Network (CNN) designs evaluated supplied dataset SGD (Stochastic Gradient Descent) optimizer. increase classification accuracy, best model fine-tuned several optimizers. It concluded that MCNN (Modified Network) achieved 99.5% accuracy F1 score 1.00 disease phase module. Enhanced GoogleNet Adam optimizer a 0.997 illnesses, which much higher than previous Thus, work adapt suggested strategy different crops identify diagnose illnesses more effectively.

Language: Английский

---

Evaluating Machine Learning Models for Rainfall Prediction: A Case Study of Nyando in Kenya

Published: Dec. 22, 2023

This paper presents a comprehensive evaluation of machine learning algorithms for rainfall prediction in the Nyando region. The study employs LSTM, XGBoost, Random Forest, and SVR algorithms, exploring both univariate multivariate models to enhance accuracy predictions. Additionally, examines three different outlier filtering methods assesses their impact on final outcomes. research endeavours contribute valuable insights field disaster management. By providing accurate reliable predictions, this aims aid communities region similar areas efforts effectively mitigate adverse impacts extreme weather events.

Language: Английский

---