A decision support system for safer airplane landings: Predicting runway conditions using XGBoost and explainable AI

Midtfjord, Alise Danielle; Bin, Riccardo De; Huseby, Arne Bang

doi:10.1016/j.coldregions.2022.103556

Cited by 17 publications

(8 citation statements)

References 41 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Moreover, AI-driven systems can forecast runway conditions and slippage probabilities [39], offering valuable support to runway operations teams. This sophisticated approach represents a notable advancement in runway management, particularly considering the rapid changes in weather conditions that require prompt and accurate decision-making [40]. By leveraging AI and integrating data-mining techniques, runway operations can be carried out with greater precision, efficiency, and safety to ensure smooth air traffic flow, even in challenging winter environments.…”

Section: Discussionmentioning

confidence: 99%

Advancing Insights into Runway De-Icing: Combining Infrared Thermography and Raman Spectroscopy to Assess Ice Melt

Charpentier,

Brassard,

Marchetti

et al. 2024

Applied Sciences

View full text Add to dashboard Cite

The “bare runway” principle aims to ensure passenger and employee safety by making runways more usable during winter conditions, allowing for easier removal of contaminants like snow and ice. Maintaining runway operations in winter is essential, but it involves considerable cost and environmental impacts. Greater knowledge about the de-icing and anti-icing performance of runway de-icing products (RDPs) optimizes operations. The ice melting test, as per the AS6170 standard, gauges the rate at which an RDP dissolves an ice mass to determine RDP effectiveness. Here, we introduce a novel integrated methodology for assessing RDP-related ice melting. We combine laboratory-based procedures with infrared thermography and Raman spectroscopy to monitor the condition of RDPs placed on ice. The plateau of maximum efficiency, marked by the most significant Raman peak intensity, corresponds to the peak minimum temperature, indicating optimal RDP performance. Beyond this point, RDP efficacy declines, and the system temperature, including melted contaminants and RDP, approaches ambient temperature. Effective RDP performance persists when the ambient temperature exceeds the mixture’s freezing point; otherwise, a freezing risk remains. The initial phases of RDP–ice contact involve exothermic reactions that generate brine, followed by heat exchange with surrounding ice to encourage melting. The final phase is complete ice melt, leaving only brine with reduced heat exchange on the surface. By quantifying these thermal and chemical changes, we gain a deeper understanding of RDP-related ice melting, and a more robust assessment can be provided to airports using RDPs.

show abstract

Section: Discussionmentioning

confidence: 99%

Advancing Insights into Runway De-Icing: Combining Infrared Thermography and Raman Spectroscopy to Assess Ice Melt

Charpentier,

Brassard,

Marchetti

et al. 2024

Applied Sciences

View full text Add to dashboard Cite

show abstract

“…We opted for the eXtreme Gradient Boosting (XGBoost) algorithm, given its established reputation for accuracy and robustness in diverse research domains (Hastie et al, 2009;Midtfjord et al, 2022;Oh et al, 2021{Chen, 2016}. XGBoost has several crucial advantages for our analysis.…”

Section: Explainable Artificial Intelligencementioning

confidence: 99%

Application of Explainable Artificial Intelligence to model the influence of firm-specific factors on stock performance in sub-Saharan Africa during COVID-19.

Ncube,

Sibanda,

Matenda

2024

Preprint

View full text Add to dashboard Cite

The COVID-19 pandemic has brought forth a novel and challenging investment environment for stock markets in both developed and emerging economies. This study delves into the under-researched field of factor investing, analysing the influence of firm-specific factors on stock performance in sub-Saharan Africa during the pandemic in four selected stock exchanges. Using explainable artificial intelligence (XAI), we analysed the cross-sectional data of firms listed on these exchanges in both the pre-pandemic and pandemic periods to identify the key factors driving stock returns. Our findings reveal that quality factors such as high profitability and low leverage emerged as the dominant drivers of resilience during the pandemic. This contrasts with the pre-pandemic scenario, where traditional factors, such as size, momentum, and value, were the primary determinants. We also find that during the pandemic, small-cap stocks generally outperformed large-cap stocks on larger stock exchanges, whereas large-cap stocks outperformed on smaller stock exchanges. The pandemic also ushered in a shift in the norm, as we found that highly volatile stocks outperformed low-volatility stocks and stocks classified as winners in the pre-pandemic period experienced a momentum crash. Further analysis explores the potential of factor-based portfolios. We demonstrate that portfolios constructed based on the identified key factors consistently outperform their market-cap weighted counterparts across all exchanges, both pre- and post-pandemic. This research highlights the importance of focusing on firm-specific factors to construct more resilient investment portfolios and potentially attain superior returns, even during times of economic turbulence.

show abstract

“…Czado and Keilegom (2022); Deresa and Keilegom (2022) proposed statistical methods for parametric survival marginals with identifiability guarantee. Midtfjord, Bin, and Huseby (2022) and Gharari et al (2023) proposed to use boosting and neural network for maximizing the log-likelihood with pre-specified copulas, respectively. To the best of our knowledge, all existing copula-based approaches require practitioners to provide the ground truth copula.…”

Section: Related Workmentioning

confidence: 99%

Deep Copula-Based Survival Analysis for Dependent Censoring with Identifiability Guarantees

Zhang,

Ling,

Zhang

2024

AAAI

View full text Add to dashboard Cite

Censoring is the central problem in survival analysis where either the time-to-event (for instance, death), or the time-to censoring (such as loss of follow-up) is observed for each sample. The majority of existing machine learning-based survival analysis methods assume that survival is conditionally independent of censoring given a set of covariates; an assumption that cannot be verified since only marginal distributions is available from the data. The existence of dependent censoring, along with the inherent bias in current estimators has been demonstrated in a variety of applications, accentuating the need for a more nuanced approach. However, existing methods that adjust for dependent censoring require practitioners to specify the ground truth copula. This requirement poses a significant challenge for practical applications, as model misspecification can lead to substantial bias. In this work, we propose a flexible deep learning-based survival analysis method that simultaneously accommodate for dependent censoring and eliminates the requirement for specifying the ground truth copula. We theoretically prove the identifiability of our model under a broad family of copulas and survival distributions. Experiments results from a wide range of datasets demonstrate that our approach successfully discerns the underlying dependency structure and significantly reduces survival estimation bias when compared to existing methods.

show abstract

A decision support system for safer airplane landings: Predicting runway conditions using XGBoost and explainable AI

Cited by 17 publications

References 41 publications

Advancing Insights into Runway De-Icing: Combining Infrared Thermography and Raman Spectroscopy to Assess Ice Melt

Advancing Insights into Runway De-Icing: Combining Infrared Thermography and Raman Spectroscopy to Assess Ice Melt

Application of Explainable Artificial Intelligence to model the influence of firm-specific factors on stock performance in sub-Saharan Africa during COVID-19.

Deep Copula-Based Survival Analysis for Dependent Censoring with Identifiability Guarantees

Contact Info

Product

Resources

About