Prediction and behavioral analysis of travel mode choice: A comparison of machine learning and logit models

Travel Behaviour and Society, Год журнала: 2020, Номер 20, С. 22 - 35

Опубликована: Фев. 18, 2020

Язык: Английский

---

Random forests for global sensitivity analysis: A selective review

Reliability Engineering & System Safety, Год журнала: 2020, Номер 206, С. 107312 - 107312

Опубликована: Ноя. 12, 2020

Язык: Английский

---

Crowdsourced air temperatures contrast satellite measures of the urban heat island and its mechanisms

Science Advances, Год журнала: 2021, Номер 7(22)

Опубликована: Май 26, 2021

Satellites overestimate urban heat islands.

Язык: Английский

---

Machine learning and deep learning—A review for ecologists

Methods in Ecology and Evolution, Год журнала: 2023, Номер 14(4), С. 994 - 1016

Опубликована: Фев. 13, 2023

Abstract The popularity of machine learning (ML), deep (DL) and artificial intelligence (AI) has risen sharply in recent years. Despite this spike popularity, the inner workings ML DL algorithms are often perceived as opaque, their relationship to classical data analysis tools remains debated. Although it is assumed that excel primarily at making predictions, can also be used for analytical tasks traditionally addressed with statistical models. Moreover, most discussions reviews on focus mainly DL, failing synthesise wealth different advantages general principles. Here, we provide a comprehensive overview field starting by summarizing its historical developments, existing algorithm families, differences traditional tools, universal We then discuss why when models prediction where they could offer alternatives methods inference, highlighting current emerging applications ecological problems. Finally, summarize trends such scientific causal ML, explainable AI, responsible AI may significantly impact future. conclude powerful new predictive modelling analysis. superior performance compared explained higher flexibility automatic data‐dependent complexity optimization. However, use inference still disputed predictions creates challenges interpretation these Nevertheless, expect become an indispensable tool ecology evolution, comparable other tools.

Язык: Английский

---

Counterfactuals and causability in explainable artificial intelligence: Theory, algorithms, and applications

Information Fusion, Год журнала: 2021, Номер 81, С. 59 - 83

Опубликована: Ноя. 13, 2021

Язык: Английский

---

Causal Interpretability for Machine Learning - Problems, Methods and Evaluation

ACM SIGKDD Explorations Newsletter, Год журнала: 2020, Номер 22(1), С. 18 - 33

Опубликована: Май 13, 2020

Machine learning models have had discernible achievements in a myriad of applications. However, most these are black-boxes, and it is obscure how the decisions made by them. This makes unreliable untrustworthy. To provide insights into decision making processes models, variety traditional interpretable been proposed. Moreover, to generate more humanfriendly explanations, recent work on interpretability tries answer questions related causality such as "Why does this model decisions?" or "Was specific feature that caused model?". In work, aim causal referred models. The existing surveys covered concepts methodologies interpretability. we present comprehensive survey from aspects problems methods. addition, provides in-depth evaluation metrics for measuring interpretability, which can help practitioners understand what scenarios each metric suitable.

Язык: Английский

---

Machine learning applications in microbial ecology, human microbiome studies, and environmental monitoring

Computational and Structural Biotechnology Journal, Год журнала: 2021, Номер 19, С. 1092 - 1107

Опубликована: Янв. 1, 2021

Advances in nucleic acid sequencing technology have enabled expansion of our ability to profile microbial diversity. These large datasets taxonomic and functional diversity are key better understanding ecology. Machine learning has proven be a useful approach for analyzing community data making predictions about outcomes including human environmental health. applied profiles been used predict disease states health, quality presence contamination the environment, as trace evidence forensics. appeal powerful tool that can provide deep insights into communities identify patterns data. However, often machine models black boxes specific outcome, with little how arrived at predictions. Complex algorithms may value higher accuracy performance sacrifice interpretability. In order leverage more translational research related microbiome strengthen extract meaningful biological information, it is important interpretable. Here we review current trends applications ecology well some challenges opportunities broad application communities.

Язык: Английский

---

Selected machine learning approaches for predicting the interfacial bond strength between FRPs and concrete

Construction and Building Materials, Год журнала: 2020, Номер 270, С. 121456 - 121456

Опубликована: Ноя. 9, 2020

Язык: Английский

---

Biohorology and biomarkers of aging: Current state-of-the-art, challenges and opportunities

Ageing Research Reviews, Год журнала: 2020, Номер 60, С. 101050 - 101050

Опубликована: Апрель 6, 2020

The aging process results in multiple traceable footprints, which can be quantified and used to estimate an organism's age. Examples of such biomarkers include epigenetic changes, telomere attrition, alterations gene expression metabolite concentrations. More than a dozen clocks use molecular features predict age, each them utilizing different data types training procedures. Here, we offer detailed comparison existing mouse human clocks, discuss their technological limitations the underlying machine learning algorithms. We also promising future directions research biohorology - science measuring passage time living systems. Overall, expect deep learning, neural networks generative approaches next power tools this timely actively developing field.

Язык: Английский

---

How Do Climate and Catchment Attributes Influence Flood Generating Processes? A Large‐Sample Study for 671 Catchments Across the Contiguous USA

Water Resources Research, Год журнала: 2021, Номер 57(4)

Опубликована: Фев. 10, 2021

Abstract Hydrometeorological flood generating processes (excess rain, short long snowmelt, and rain‐on‐snow) underpin our understanding of behavior. Knowledge about improves hydrological models, frequency analysis, estimation climate change impact on floods, etc. Yet, not much is known how catchment attributes influence the spatial distribution processes. This study aims to offer a comprehensive structured approach close this knowledge gap. We employ large sample (671 catchments across contiguous United States) evaluate use two complementary approaches: A statistics‐based which compares attribute distributions different processes; random forest model in combination with an interpretable machine learning (accumulated local effects [ALE]). The ALE method has been used often hydrology, it overcomes significant obstacle many statistical methods, confounding effect correlated attributes. As expected, we find (fraction snow, aridity, precipitation seasonality, mean precipitation) be most influential process distribution. However, varies both type. also can predicted for ungauged relatively high accuracy ( R 2 between 0.45 0.9). implication these findings should considered future studies, as changes characteristics

Язык: Английский

---