An Anomaly Detection Method of Industrial Data Based on Stacking  Integration

Wang, Kunkun; Liu, Xianda

doi:10.32604/jai.2021.016706

Cited by 7 publications

(5 citation statements)

References 6 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…EDT is a combination of several DT whereas bagging formulates x number of bags, containing a subset of the original data which is further trained. The predicted outcomes from each bags determine the mode of all results for classification [41]. Adaboost utilizes a set of weights over the training set and minimizes the weighted error after each iteration to formulate multiple hypotheses.…”

Section: The Dataset and Methodologymentioning

confidence: 99%

“…The stacked ensemble (SE) learning was first demonstrated by Wolpert [40], where authors exhibited a stacked formation of strong and weak ML algorithms which could improvise framework accuracy and filter samples by boosting training competency and lessening overfitting issues. Several researchers have represented their approaches based on SE classification for numerous areas such as network intrusion detection (NID), statistics, forecasting big data, and human activity recognition [40,41]. The primary concept of SE architecture is the combination of weak ML algorithms to generate a strong framework.…”

Section: Stacked Ensemble (Se) Classifiermentioning

confidence: 99%

“…The base learners firstly forecast sensed information on the primary level. The prediction output from these candidates are then recognized and combined as a new training set to the second level classifier [41]. The meta learner thereafter, computes the loss values of the base algorithms before generating the last prediction [40].…”

Section: Stacked Ensemble (Se) Classifiermentioning

confidence: 99%

“…Support Vector Machines (SVM) is a statistical-based ML algorithm applied for classification and regression problems [41,43]. The prime conception of SVM is to obtain hyperplane (decision boundary) by maximizing the margin of the sample points known as support vectors, leading to separating different classes [43].…”

Section: Support Vector Machinementioning

confidence: 99%

“…Ensemble learning, a combination of multiple ML algorithms, enhances the overall performances of a framework [40,41]. In this paper, EDT, employing for SEC to detect movement and breathing rate; bagging, Adaboost and RUSboost are applied, respectively, for breathing pattern recognition using Bayesian optimization.…”

Section: Ensemble Treementioning

confidence: 99%

See 4 more Smart Citations

Smart factory floor safety monitoring using UWB sensor

Islam

Lee

Kim

2022

IET Science Measure & Tech

View full text Add to dashboard Cite

Chemical asphyxiation at petrochemical factories can provoke the unconsciousness or death of factory workers through suffocation. Some chemicals vaporize and mix with air without showing any warning properties that raise the risk of oxygen deficiency. In light of this, Industry 5.0 focuses more on human‐centricity than technology‐driven implementations to ensure secured and work‐friendly environments in industries. Recently, research on factory safety management dependent on the Internet of things (IoT) sensors have been executed unwaveringly. In this work, the ultra‐wideband (UWB) sensor is adopted to recognize the motion and breathing pattern of workers in smart factory scenarios. After capturing the data from the UWB sensor in real‐time, the proposed dataset is further inspected by the deep learning (DL) and traditional machine learning (ML) approaches. Twofold detection schemes are considered where the movement and vital patterns are distinguished first by the stacked ensemble (SE) and the long short‐term memory (LSTM) frameworks. The Bayesian optimized ensemble learning (EL) and bidirectional (Bi‐LSTM) models are further occupied to analyze abnormalities in the breathing rate of a worker in the smart shop floors. The investigated outcome shows that the DL frameworks (LSTM and Bi‐LSTM) outperformed the others by acquiring 99.90% and 99.94% accuracy in 147 s and 293 s, respectively. The devised perception indicates prominent attainment to the smart factory shop floor, Internet of medical things (IoMT), the smart city paradigm, and e‐health appliances.

show abstract

Section: The Dataset and Methodologymentioning

confidence: 99%

Section: Stacked Ensemble (Se) Classifiermentioning

confidence: 99%

Section: Stacked Ensemble (Se) Classifiermentioning

confidence: 99%

Section: Support Vector Machinementioning

confidence: 99%

Section: Ensemble Treementioning

confidence: 99%

See 3 more Smart Citations

Smart factory floor safety monitoring using UWB sensor

Islam

Lee

Kim

2022

IET Science Measure & Tech

View full text Add to dashboard Cite

show abstract

Design of Intelligent Integration System for Multi-source Industrial Field Data Based on Machine Learning

Zhuo,

Kang

2024

Lecture Notes of the Institute for Computer Sciences, Social Informatics and Telecommunications Engineering

View full text Add to dashboard Cite

Lithology identification based on interpretability integration learning

Lin¹,

Yin

2023

Earth Sci Inform

View full text Add to dashboard Cite

A lithology intelligent identi cation interpretability model is proposed based on Ensemble LearningStacking, Permutation Importance (PI) and Local Interpretable Model-agnostic Explanations (LIME). The method aiming to provide more accurate geological information and more scienti c theoretical support for oil and gas resource exploration. Two logging datasets from the public domain were used as experiments, and support vector machine (SVM), random forest (RF) and naive bayes (NB) were used as primary learners, and SVM as secondary learners, to classify lithology through stacking algorithm. Then, the evaluation indexes such as Area Under Curve (AUC), precision, recall and F1-score were used to verify its accuracy, and PI and LIME were used to explain the lithology identi cation model. The study shows that the results of the stacking algorithm have the best indexes and the highest prediction accuracy. In terms of overall interpretation, PHIND, GR and RT have the most in uence on lithology identi cation of a natural gas protection area in the United States; DEN, CAL and PEF have the most in uence on lithology identi cation in Daqing Oil eld in China. Interpreted from the perspective of a single sample, the LIME algorithm is able to give a quantitative prediction probability and the degree of in uence of the characteristic variables.

show abstract

An Anomaly Detection Method of Industrial Data Based on Stacking Integration

Cited by 7 publications

References 6 publications

Smart factory floor safety monitoring using UWB sensor

Smart factory floor safety monitoring using UWB sensor

Design of Intelligent Integration System for Multi-source Industrial Field Data Based on Machine Learning

Lithology identification based on interpretability integration learning

Contact Info

Product

Resources

About