Real-time data visual monitoring of triboelectric nanogenerators enabled by Deep learning

Nano Energy, Journal Year: 2024, Volume and Issue: 130, P. 110186 - 110186

Published: Aug. 27, 2024

Language: Английский

---

Avoiding common machine learning pitfalls

Patterns, Journal Year: 2024, Volume and Issue: 5(10), P. 101046 - 101046

Published: Aug. 28, 2024

The bigger pictureMachine learning has transitioned from a niche pursuit to one with mass appeal. Thanks the accessibility of modern machine tools, it is now very easy get started in learning, yet this ease use masks underlying complexities doing learning. This, coupled relatively inexperienced community practitioners, led flawed practices, which are reflected issues such as poor reproducibility within machine-learning-based studies.This tutorial aims address problem by educating practitioners about many things that can go wrong when applying and providing guidance on how avoid these pitfalls. However, just part longer-term process needed improve practice, will only meet its ambitions if able become robust trusted applied discipline. Other factors have role play include better standardization, regulation.SummaryMistakes practice commonplace result loss confidence findings products This outlines common mistakes occur using what be done them. While should accessible anyone basic understanding techniques, focuses particular concern academic research, need make rigorous comparisons reach valid conclusions. It covers five stages process: do before model building, reliably build models, robustly evaluate compare models fairly, report results.

Language: Английский

---

A Transformer Oil Temperature Prediction Method Based on Data-Driven and Multi-Model Fusion

Processes, Journal Year: 2025, Volume and Issue: 13(2), P. 302 - 302

Published: Jan. 22, 2025

A power transformer is an important part of the system, and oil temperature state parameter that reflects operation transformer. The accurate prediction can ensure safe stable Given lack a practical effective data processing process problem most current research conducted on small-scale ideal datasets, this paper proposes method based data-driven multi-model fusion. first analyses processes actual inspection data; it then uses fusion to model predict temperature. base was trained using machine learning method, secondary improved TSSA-BP neural network. sparrow search algorithm (TSSA) used optimise parameters BP network improve convergence accuracy performance model. are classified according cooling mode, operating voltage, other indicators, eight groups experimental datasets under different conditions constructed for modelling prediction. results show maximum root mean square error absolute percentage 1.0877 1.58%, compared with methods, better than which verifies practicability feasibility predicting data.

Language: Английский

---

Transformative strategies in photocatalyst design: merging computational methods and deep learning

Journal of Materials Informatics, Journal Year: 2024, Volume and Issue: 4(4)

Published: Dec. 31, 2024

Photocatalysis is a unique technology that harnesses solar energy through in-situ processes, operating without the need for external inputs. It integral to advancing environmental, energy, chemical, and carbon-neutral objectives, promoting dual goals of pollution control carbon reduction. However, conventional approach photocatalyst design faces challenges such as inefficiency, high costs, low success rates, highlighting integrating modern technologies seeking new paradigms. Here, we demonstrate comprehensive overview transformative strategies in design, combining computational materials science with deep learning technologies. The review covers fundamental principles followed by examination methods workflow deep-learning-assisted design. Deep approaches are extensively reviewed, focusing on discovery novel photocatalysts, microstructure property optimization, approaches, application exploration, mechanistic insights into photocatalysis. Finally, highlight synergy between multidimensional computation learning, while discussing future directions development. This offers summary offering not only enhance development photocatalytic but also expand practical applications photocatalysis various domains.

Language: Английский

---

Forecasting Follies: Machine Learning from Human Errors

Journal of risk and financial management, Journal Year: 2025, Volume and Issue: 18(2), P. 60 - 60

Published: Jan. 28, 2025

Reliable inflation forecasts are essential for both business operations and macroeconomic policy making. This study explores the potential of using machine learning (ML) techniques to improve accuracy human inflation. Specifically, we develop examine ML-centered forecast adjustment procedures where advanced ML employed predict thus mitigate errors forecasts, akin how an AI-powered spell grammar checker helps prevent mistakes in writing. Our empirical exercises demonstrate benefits several popular techniques, such as elastic net, LASSO, ridge regressions, provide evidence their ability our own benchmark those reported by frequent participants US Survey Professional Forecasters. The proposed this paper conceptually appealing, widely applicable, empirically effective reducing bias improving accuracy.

Language: Английский

---

Issues and trends in generative AI technologies for decision making

Intelligent Decision Technologies, Journal Year: 2025, Volume and Issue: unknown

Published: Feb. 24, 2025

Generative AI (GenAI) technologies are examined through the lens of issues and trends related to decision making. After examining foundations technology particularly large language models (LLM), opportunities for GenAI be used in decision-making process intelligence, design, choice implementation explored. With its ability rapidly generate insights, present optimized solutions, provide detailed analysis given input, has demonstrated that it can assist augment human Although systems have potential transform content creation cognition, they also raise around accuracy, misinformation, ethics, bias, morality, social impacts, privacy, copyright, legality, explainability, among others. Addressing these challenges is important maximize efficacy

Language: Английский

---

Unmanned Aerial Vehicle-Based Hyperspectral Imaging Integrated with a Data Cleaning Strategy for Detection of Corn Canopy Biomass, Chlorophyll, and Nitrogen Contents at Plant Scale

Remote Sensing, Journal Year: 2025, Volume and Issue: 17(5), P. 895 - 895

Published: March 3, 2025

The high-frequency detection of plant-scale crop growth in the field has great significance for achieving precise management and improving breeding practices. In this study, biomass (BM), chlorophyll (Chl), total nitrogen (TN) contents upper three leaves corn canopy are taken as examples, unmanned aerial vehicle (UAV) indoor hyperspectral imaging (HSI) models established using partial least squares regression support vector machine regression, respectively. performance UAV HSI model was notably lower comparison to model. Therefore, a data cleaning strategy integrated with RGB image information is further proposed, which involves eliminating points serious interference from non-related plant. After cleaning, R2C BM, Chl, TN detected through reached 0.537, 0.852, 0.657, representing an improvement over 70%. RMSEP values were low 0.50 g, 2.2 SPAD, 0.258%, comparable those obtained This study demonstrates that proposed can enable rapid leaf properties at plant scale field, supporting characterization parameters field.

Language: Английский

---

Estimation of PM_2.5 Concentration in DKI Jakarta from Sentinel-5P Imagery by Considering Meteorological Factors Using Random Forest Approach

BIO Web of Conferences, Journal Year: 2025, Volume and Issue: 167, P. 03003 - 03003

Published: Jan. 1, 2025

Poor air quality, caused by high pollutant levels, harms the environment and public health. Fine particulate matter (PM 2.5 ), less than μm in diameter, is a major concern quality observations due to its ability penetrate respiratory system, increasing risks of lung cancer, premature death, unnatural births. Jakarta faces severe pollution, yet monitoring network remains limited. To address this, this study employs machine learning, specifically random forest algorithms, using spatial regression model PM levels. The variables used are meteorological elements particulates gasses obtained utilizing remote sensing. It was found that R 2 value 0.793 implies accuracy reaches 79.3 percent RMSE 8.28 μg/m3. pattern formed modelling follows rainy season dry season, where highest parameter JJA month (June, July August), finally at lowest DJF (December, January February).

Language: Английский

---

Detection of Pedestrian Movement Poses in High-Speed Autonomous Driving Environments Using DVS

Communications in computer and information science, Journal Year: 2025, Volume and Issue: unknown, P. 54 - 66

Published: Jan. 1, 2025

Language: Английский

---

Explanation and elaboration of MedinAI: guidelines for reporting artificial intelligence studies in medicines, pharmacotherapy, and pharmaceutical services

International Journal of Clinical Pharmacy, Journal Year: 2025, Volume and Issue: unknown

Published: April 18, 2025

Language: Английский

---