相关参考文献
注意：仅列出部分参考文献，下载原文获取全部文献信息。
Article
Computer Science, Artificial Intelligence
Ryan B. Christianson et al.
Summary: The canonical technique for nonlinear modeling of spatial/point-referenced data is kriging in geostatistics and Gaussian Process (GP) regression for surrogate modeling and statistical learning. This article compares kriging and GPs, highlighting their similarities and differences. GPs offer automated kernel/variogram inference and scalable handling of large training data, making them more economical, accurate, and capable of better uncertainty quantification compared to classic kriging. The article also demonstrates the capability of GPs in accommodating left-censoring of small measurements in mining data and borehole assays.
STATISTICAL ANALYSIS AND DATA MINING
(2023)
Article
Energy & Fuels
Auref Rostamian et al.
Summary: This study utilized various ML models for well log data prediction, making breakthroughs in feature selection and hyperparameter optimization. Results showed that RF, k-NN, and XGBoost performed excellently. Preserving all data in the models can enhance prediction performance, while reducing the number of features may lead to performance deterioration.
JOURNAL OF PETROLEUM SCIENCE AND ENGINEERING
(2022)
Article
Geosciences, Multidisciplinary
Sandra De Iaco et al.
Summary: In recent years, there has been a steady growth in the number of papers applying machine learning methods in the earth sciences. Despite their different origins, machine learning and geostatistics share concepts and methods. Machine learning, with its focus on algorithms and ability to uncover hidden structures in big data sets, provides new tools for exploration and prediction in the earth sciences. On the other hand, geostatistics offers interpretable models of spatial dependence. This special issue aims to investigate the applications of machine learning methods and hybrid approaches combining machine learning and geostatistics, advancing our understanding and predictive ability of spatial processes.
MATHEMATICAL GEOSCIENCES
(2022)
Article
Multidisciplinary Sciences
Moaz Hiba et al.
Summary: In this study, an artificial neural network model is used to predict drilling parameters, showing higher accuracy and reliability compared to existing methods.
ARABIAN JOURNAL FOR SCIENCE AND ENGINEERING
(2022)
Article
Energy & Fuels
Lei Wu et al.
Summary: This study effectively predicted well-logging data by combining CNN and LSTM neural networks, utilizing the PSO algorithm to determine optimal hyperparameters of the structure.
Article
Energy & Fuels
Nelson R. K. Tatsipie et al.
Summary: This study demonstrated the successful use of deep learning models to generate synthetic logs for a set of 40 wells in the upper Bakken using well log data from 97 wells. By utilizing petroleum engineering concepts and RNNs, the data-driven models proved effective in creating the remaining logs and maximizing the value of a gamma ray log. With the ability to generate logs of similar quality for all wells in a shale asset starting only with gamma ray data, superior reservoir evaluation and improved reservoir management can be achieved.
JOURNAL OF PETROLEUM SCIENCE AND ENGINEERING
(2021)
Article
Energy & Fuels
Lucas Abreu Blanes de Oliveira et al.
Summary: Geochemical logs are essential for hydrocarbon reservoir characterization, but in reduced logging operations in Brazilian pre-salt carbonate reservoirs, the geochemical tool is no longer acquired for cost reduction. This research develops synthetic geochemical logs using machine learning algorithms that can substitute actual logs with high confidence, providing quality data for formation evaluation. Robust models are trained for Al, Ca, Fe, Mg, Si, S, Ti, and Na, with R-2 above 0.70, and RMSE between 10(-2) to 10(-4), showing good agreement with acquired logs and general trends of the pre-salt formations.
JOURNAL OF PETROLEUM SCIENCE AND ENGINEERING
(2021)
Article
Computer Science, Interdisciplinary Applications
Tao Cui et al.
Summary: Gaussian processes provide statistically optimal predictions with unbiasedness and precision, yet their implementation in environmental science is less routine. By learning from multiple data sources such as AEM and DEM data, GPs have demonstrated flexibility and capability in generating groundwater salinity maps with rich local details and quantified uncertainty to support decision making. The use of nonpoint data with regional spatial coverage is valuable for providing realistic heterogeneity in aquifer properties, especially when point measurements are sparse and multiple predictors are available.
ENVIRONMENTAL MODELLING & SOFTWARE
(2021)
Article
Computer Science, Artificial Intelligence
Ahmed Gowida et al.
Summary: The study developed three artificial intelligence models to predict UCS of downhole formations in real-time, with an optimized UCS correlation outperforming existing correlations. The ANN model showed the highest accuracy with a correlation coefficient of 0.99 and an average absolute percentage error of 3.48%, demonstrating the effectiveness of AI in predicting UCS values.
NEURAL COMPUTING & APPLICATIONS
(2021)
Article
Computer Science, Information Systems
Xu Yu et al.
Summary: This study introduces a Two-Sided CDCF model based on Selective Ensemble learning to balance recommendation accuracy and efficiency. By combining user-sided and item-sided auxiliary domains, converting CDCF problems to ensemble learning problems, and solving a bi-objective optimization problem, the model achieves Pareto optimal solutions for the selective ensemble learning problem.
INFORMATION PROCESSING & MANAGEMENT
(2021)
Article
Energy & Fuels
Timur Merembayev et al.
Summary: The study evaluates lithofacies values using machine learning algorithms with data from Kazakhstan and Norway, with Random Forest model scoring the highest among the considered algorithms. The results are consistent with the SHapley Additive exPlanations (SHAP) framework.
Article
Chemistry, Multidisciplinary
Hany Gamal et al.
Summary: This study utilized machine learning techniques to predict ECD based on drilling parameters, constructed models, and validated them through key performance indices. The results showed that both models performed well in ECD estimation application.
Article
Geochemistry & Geophysics
Yanxiang Yu et al.
Summary: DTC and DTS logs are crucial for subsurface characterization and seismic-well tie, but are often missing in oil and gas wells. Machine-learning methods can be used to predict these logs in training models. Data cleaning and clustering are critical for improving model performance.
Article
Chemistry, Multidisciplinary
Mahmoud Desouky et al.
Summary: This research utilized machine learning to generate simple and accurate propped fracture conductivity correlations in unconventional gas shale formations. The artificial neural network (ANN) model provided the highest accuracy in fracture conductivity estimation. Easily obtained input parameters were identified as crucial for predicting fracture conductivity, with closure stress being the most significant parameter.
Article
Engineering, Marine
Fahimeh Hadavimoghaddam et al.
Summary: This study successfully predicted water saturation using new machine learning algorithms, avoiding the issues of traditional methods relying on resistivity log data in specific formations. Super Learner and XGBoost were found to produce the most accurate predictions, with Super Learner considered the best among all models.
JOURNAL OF MARINE SCIENCE AND ENGINEERING
(2021)
Article
Energy & Fuels
Mohammad Islam Miah et al.
JOURNAL OF PETROLEUM SCIENCE AND ENGINEERING
(2020)
Article
Chemistry, Multidisciplinary
Touhid Mohammad Hossain et al.
APPLIED SCIENCES-BASEL
(2020)
Article
Energy & Fuels
David Onalo et al.
JOURNAL OF PETROLEUM EXPLORATION AND PRODUCTION TECHNOLOGY
(2020)
Article
Computer Science, Artificial Intelligence
Xu Yu et al.
PATTERN RECOGNITION
(2019)
Review
Physics, Multidisciplinary
Claudio Zeni et al.
ADVANCES IN PHYSICS-X
(2019)
Article
Mathematics, Interdisciplinary Applications
Eric Schulz et al.
JOURNAL OF MATHEMATICAL PSYCHOLOGY
(2018)
Article
Computer Science, Artificial Intelligence
Xu Yu et al.
KNOWLEDGE-BASED SYSTEMS
(2018)
Article
Energy & Fuels
Solomon Asante-Okyere et al.
Review
Computer Science, Artificial Intelligence
Amit Saxena et al.
Article
Energy & Fuels
M. Asoodeh et al.
ENERGY SOURCES PART A-RECOVERY UTILIZATION AND ENVIRONMENTAL EFFECTS
(2015)
Article
Energy & Fuels
Fouad Bahrpeyma et al.
JOURNAL OF PETROLEUM SCIENCE AND ENGINEERING
(2013)
Article
Computer Science, Artificial Intelligence
J. Bobadilla et al.
KNOWLEDGE-BASED SYSTEMS
(2013)
