☆ 4.8 Article

Augmenting energy time-series for data-efficient imputation of missing values

APPLIED ENERGY (2023)

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 Construction & Building Technology

Implications of properties and quality of indoor sensor data for building machine learning applications: Two case studies in smart campuses

Miia Lillstrang et al.

Summary: This article presents a case study on two real-life data sets from university campus buildings located in different climates, applying different sensor network settings. The study aims to build understanding of the environment in which machine learning models utilizing sensor data operate and provides information about the properties and quality aspects of public building data.

BUILDING AND ENVIRONMENT (2022)

添加到收藏夹

Article Construction & Building Technology

Impacts of data preprocessing and selection on energy consumption prediction model of HVAC systems based on deep learning

Ziwei Xiao et al.

Summary: This paper studies the impact of data processing methods and feature selection on HVAC energy consumption prediction using data-driven models. By developing long and short-term memory models based on historical data, the effects of two data smoothing methods, feature selection, training set volumes, and model update frequency are analyzed and compared to improve prediction accuracy.

ENERGY AND BUILDINGS (2022)

添加到收藏夹

Article Construction & Building Technology

A guideline to document occupant behavior models for advanced building controls

Bing Dong et al.

Summary: This paper proposes a comprehensive and standardized guideline for occupant-behavior model documentation. Through literature screening and studying existing models, it identifies practices and gaps in various phases. The research indicates a lack of uniformity in current model documentation.

BUILDING AND ENVIRONMENT (2022)

添加到收藏夹

Article Chemistry, Multidisciplinary

An Improved Method Based on EEMD-LSTM to Predict Missing Measured Data of Structural Sensors

Zengshun Chen et al.

Summary: This study proposes a hybrid method combining ensemble empirical mode decomposition (EEMD) and deep learning models for predicting missing measured signal data. The method is validated using shaking-table experiments on a liquefied natural gas (LNG) storage tank. The results demonstrate the superior complementary performance of the proposed method and provide suggestions for using the EEMD-LSTM model under different conditions.

APPLIED SCIENCES-BASEL (2022)

添加到收藏夹

Review Energy & Fuels

Transfer learning for smart buildings: A critical review of algorithms, applications, and future perspectives

Giuseppe Pinto et al.

Summary: This study provides a comprehensive overview of the applications of transfer learning in smart buildings, classifying and analyzing 77 papers. The research identifies four main application areas of transfer learning: building load prediction, occupancy detection and activity recognition, building dynamics modeling, and energy systems control.

ADVANCES IN APPLIED ENERGY (2022)

添加到收藏夹

Article Engineering, Multidisciplinary

A hybrid method coupling empirical mode decomposition and a long short-term memory network to predict missing measured signal data of SHM systems

Linchao Li et al.

Summary: A hybrid method combining empirical mode decomposition with long short-term memory deep learning network is proposed for the recovery of measured signal data in structural health monitoring systems. The method converts the missing data imputation task into a time series prediction task and utilizes a divide and conquer strategy. Experimental results demonstrate that the proposed hybrid method exhibits excellent performance in recovering measured signal data.

STRUCTURAL HEALTH MONITORING-AN INTERNATIONAL JOURNAL (2021)

添加到收藏夹

Article Energy & Fuels

Missing data imputation using mixture factor analysis for building electric load data

Dongyeon Jeong et al.

Summary: The study introduces a mixture factor analysis method for estimating missing values in building electric load data. Due to quality issues in building electric load data, a novel data imputation model is proposed to represent patterns and their cyclic rotations, providing better handling of missing data problems and improving efficiency and accuracy in model selection.

APPLIED ENERGY (2021)

添加到收藏夹

Article Thermodynamics

Data augmentation strategy for short-term heating load prediction model of residential building

Yakai Lu et al.

Summary: Data-driven models are widely used for short-term heating load prediction of buildings, but insufficient data often affects prediction performance. This study proposes a data augmentation strategy using calibrated simulation model and transfer learning-based data fusion method to improve prediction accuracy under limited data conditions, showing a significant reduction in prediction errors.

ENERGY (2021)

添加到收藏夹

Article Engineering, Electrical & Electronic

MuSDRI: Multi-Seasonal Decomposition Based Recurrent Imputation for Time Series

Yujue Zhou et al.

Summary: This paper investigates different paradigms of time series imputation methods and proposes a new learning paradigm, MuSDRI, which combines STL with RNNs to capture temporal dynamics from both long-term and short-term perspectives. Experimental results show that imputation methods considering both long-term and short-term dynamics achieve higher accuracy, with MuSDRI generally performing the best.

IEEE SENSORS JOURNAL (2021)

添加到收藏夹

Article Chemistry, Analytical

Deep Convolutional and LSTM Networks on Multi-Channel Time Series Data for Gait Phase Recognition

David Kreuzer et al.

Summary: As society ages, the prevalence of gait disorders increases, impacting quality of life and fall risks. Early and systematic diagnostic treatment of gait disorders can reduce patient suffering. While modern gait analysis systems are costly, low-cost systems like IMUs offer potential for automatic real-time motion analysis.

SENSORS (2021)

添加到收藏夹

Article Construction & Building Technology

Indoor environment data time-series reconstruction using autoencoder neural networks

Antonio Liguori et al.

Summary: This study presents a data-driven approach to fill missing data in building operation, training three different autoencoder neural networks to reconstruct short-term indoor environment data time-series. The models outperform classic numerical approaches, resulting in reconstructing the corresponding variables with average RMSEs of 0.42 degrees C, 1.30 % and 78.41 ppm.

BUILDING AND ENVIRONMENT (2021)

添加到收藏夹

Article Chemistry, Multidisciplinary

From Nearly Zero Energy to Carbon-Neutral: Case Study of a Hospitality Building

Francesco Causone et al.

Summary: This paper presents a definition of carbon-neutral building (CNB) based on available standards and scientific literature on life cycle assessment (LCA), and highlights the limitations of achieving carbon neutrality at the single urban building level in a case study. It suggests that carbon neutrality may be better approached at larger scales, such as city, regional, or national, where overarching policies can be defined.

APPLIED SCIENCES-BASEL (2021)

添加到收藏夹

Article Computer Science, Artificial Intelligence

A fractal interpolation approach to improve neural network predictions for difficult time series data

Sebastian Raubitzek et al.

Summary: This study investigates the application of deep learning methods in predicting time series data, finding that for insufficient data sets, the fractal interpolation method outperforms the linear interpolation method, generating more fine-grained time series data.

EXPERT SYSTEMS WITH APPLICATIONS (2021)

添加到收藏夹

Article Engineering, Electrical & Electronic

A Two-Stage Deep Autoencoder-Based Missing Data Imputation Method for Wind Farm SCADA Data

Xin Liu et al.

Summary: This paper introduces a novel method for imputing missing SCADA data of wind turbines with high accuracy using deep nonparametric models and a gradient-based optimization algorithm; the method is optimized in two stages and demonstrates good performance on considered benchmarks.

IEEE SENSORS JOURNAL (2021)

添加到收藏夹

Review Energy & Fuels

A Review on Data Preprocessing Techniques Toward Efficient and Reliable Knowledge Discovery From Building Operational Data

Cheng Fan et al.

Summary: This article provides a comprehensive review of the importance of data preprocessing in analyzing massive building operational data and the applications of various techniques, as well as proposing the latest data science techniques to address data challenges in the building field.

FRONTIERS IN ENERGY RESEARCH (2021)

添加到收藏夹

Article Computer Science, Artificial Intelligence

Model complexity of deep learning: a survey

Xia Hu et al.

Summary: This paper provides a systematic overview of the latest studies on model complexity in deep learning, categorizing it into expressive capacity and effective model complexity. The paper reviews existing research on these two categories based on four important factors and discusses the applications of deep learning model complexity. Finally, the paper proposes several interesting future directions.

KNOWLEDGE AND INFORMATION SYSTEMS (2021)

添加到收藏夹

Proceedings Paper Computer Science, Artificial Intelligence

A Transformer-based Framework for Multivariate Time Series Representation Learning

George Zerveas et al.

Summary: A novel framework for multivariate time series representation learning based on transformer encoder architecture is proposed, including an unsupervised pre-training scheme that outperforms currently available methods, leading to improved performance in regression and classification tasks, even with datasets containing few samples.

KDD '21: PROCEEDINGS OF THE 27TH ACM SIGKDD CONFERENCE ON KNOWLEDGE DISCOVERY & DATA MINING (2021)

添加到收藏夹

Article Computer Science, Information Systems

A GAN-Enhanced Ensemble Model for Energy Consumption Forecasting in Large Commercial Buildings

Danlan Wu et al.

Summary: Building an accurate and robust energy consumption prediction model is crucial for smart building energy management systems. This paper presents a GAN-enhanced ensemble model for energy consumption forecasting in large commercial buildings, which has been shown to significantly reduce prediction errors in experiments.

IEEE ACCESS (2021)

添加到收藏夹

Article Energy & Fuels

A Novel Ensemble Algorithm for Solar Power Forecasting Based on Kernel Density Estimation

Mohamed Lotfi et al.

ENERGIES (2020)

添加到收藏夹

Article Construction & Building Technology

Sustainable subway indoor air quality monitoring and fault-tolerant ventilation control using a sparse autoencoder-driven sensor self-validation

Jorge Loy-Benitez et al.

SUSTAINABLE CITIES AND SOCIETY (2020)

添加到收藏夹

Article Chemistry, Analytical

Bagging Ensemble of Multilayer Perceptrons for Missing Electricity Consumption Data Imputation

Seungwon Jung et al.

SENSORS (2020)

添加到收藏夹

Article Energy & Fuels

Leveraging Demand Flexibility by Exploiting Prosumer Response to Price Signals in Microgrids

Francesco Simmini et al.

ENERGIES (2020)

添加到收藏夹

Article Construction & Building Technology

A bi-directional missing data imputation scheme based on LSTM and transfer learning for building energy data

Jun Ma et al.

ENERGY AND BUILDINGS (2020)

添加到收藏夹

Article Construction & Building Technology

Anomaly detection based on machine learning in IoT-based vertical plant wall for indoor climate control

Yu Liu et al.

BUILDING AND ENVIRONMENT (2020)

添加到收藏夹

Review Multidisciplinary Sciences

Array programming with NumPy

Charles R. Harris et al.

NATURE (2020)

添加到收藏夹

Article Automation & Control Systems

End-to-End Incomplete Time-Series Modeling From Linear Memory of Latent Variables

Qianli Ma et al.

IEEE TRANSACTIONS ON CYBERNETICS (2020)

添加到收藏夹

Article Computer Science, Information Systems

SSIM-A Deep Learning Approach for Recovering Missing Time Series Sensor Data

Yi-Fan Zhang et al.

IEEE INTERNET OF THINGS JOURNAL (2019)

添加到收藏夹

Article Computer Science, Artificial Intelligence

Spatio-temporal neural networks for space-time data modeling and relation discovery

Edouard Delasalles et al.

KNOWLEDGE AND INFORMATION SYSTEMS (2019)

添加到收藏夹

Article Energy & Fuels

Analytical investigation of autoencoder-based methods for unsupervised anomaly detection in building energy data

Cheng Fan et al.

APPLIED ENERGY (2018)

添加到收藏夹

Article Construction & Building Technology

Time series distance-based methods for non-intrusive load monitoring in residential buildings

Kaustav Basu et al.

ENERGY AND BUILDINGS (2015)

添加到收藏夹

Article Biochemical Research Methods

MissForest-non-parametric missing value imputation for mixed-type data

Daniel J. Stekhoven et al.

BIOINFORMATICS (2012)

添加到收藏夹

Article Computer Science, Artificial Intelligence

Image denoising via sparse and redundant representations over learned dictionaries

Michael Elad et al.

IEEE TRANSACTIONS ON IMAGE PROCESSING (2006)

添加到收藏夹

Article Biochemical Research Methods

Missing value estimation methods for DNA microarrays

O Troyanskaya et al.

BIOINFORMATICS (2001)

添加到收藏夹

© Peeref 2019-2024. All rights reserved.