A robust correlation analysis framework for imbalanced and dichotomous data with uncertainty

Lai, Chun Sing; Tao, Yingshan; Xu, Fangyuan; Ng, Wing W. Y.; Jia, Youwei; Yuan, Haoliang; Huang, Chao; Zhao, Xu; Locatelli, Giorgio

doi:10.1016/j.ins.2018.08.017

Cited by 60 publications

(29 citation statements)

References 45 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Consequently, each considered condition has been experimentally evaluated under four different operating regimes; fifty measurements were performed for each condition. It must be noticed that the data imbalance represents a critical issue in most data science challenges and, in particular dealing with electromechanical systems, since the normal operating condition is the most represented usually [21][22]. However, the proposed methodology avoids the non-desired effects of such situation since the adaptive nature of the selforganizing maps technique considered in the proposed methodology characterizes topological aspects of the data manifold as data density, independently of the data number.…”

Section: Methodsmentioning

confidence: 99%

Industrial Data-Driven Monitoring Based on Incremental Learning Applied to the Detection of Novel Faults

Saucedo-Dorantes

Delgado-Prieto

Osornio-Ríos

et al. 2020

IEEE Trans. Ind. Inf.

View full text Add to dashboard Cite

The detection of uncharacterized events during electromechanical systems operation represents one of the most critical data challenges dealing with condition-based monitoring under the Industry 4.0 framework. Thus, the detection of novelty conditions and the learning of new patterns are considered as mandatory competencies in modern industrial applications. In this regard, this study proposes a novel multi-fault detection and identification scheme, based on machine learning, information data-fusion, novelty-detection and incremental learning. First, statistical time-domain features estimated from multiple physical magnitudes acquired from the electrical motor under inspection are fused under a feature-fusion level scheme. Second, a selforganizing map structure is proposed to construct a data-based model of the available conditions of operation. Third, the incremental learning of the condition-based monitoring scheme is performed adding self-organizing structures and optimizing their projections through a linear discriminant analysis. The performance of the proposed scheme is validated under a complete set of experimental scenarios and results compared with a classical approach.

show abstract

Section: Methodsmentioning

confidence: 99%

Industrial Data-Driven Monitoring Based on Incremental Learning Applied to the Detection of Novel Faults

Saucedo-Dorantes

Delgado-Prieto

Osornio-Ríos

et al. 2020

IEEE Trans. Ind. Inf.

View full text Add to dashboard Cite

show abstract

“…In our study, it is found that load values of different years on the same date (LDYSD), such as the daily peak load of 1 December 2014, and that of 1 December 2015, have a strong correlation based on correlation analysis [29]. The discovery indicates that the corresponding historical loads have some implicit relationships with the load to be predicted.…”

Section: Introductionmentioning

confidence: 93%

Load forecasting based on deep neural network and historical data augmentation

Lai

Wang

et al. 2020

IET Generation, Transmission & Distribution

Self Cite

View full text Add to dashboard Cite

Load forecasting is a complex non-linear problem with high volatility and uncertainty. This paper presents a novel load forecasting method known as deep neural network and historical data augmentation (DNN-HDA). The method utilizes HDA to enhance regression by DNN for monthly load forecasting, considering that the historical data to have a high correlation with the corresponding predicted data. To make the best use of the historical data, one year's historical data is combined with the basic features to construct the input vector for a predicted load. In this way, if there is C years' historical data, one predicted load can have C input vectors to create the same number of samples. DNN-HDA increases the number of training samples and enhances the generalization of the model to reduce the forecasting error. The proposed method is tested on daily peak loads from 2006 to 2015 of Austria, Czech and Italy. Comparisons are made between the proposed method and several state-of-the-art models. DNN-HDA outperforms DNN by 44%, 38% and 63% on the three datasets, respectively.

show abstract

“…Following Pearson correlation coefficient rules [24], the value rule and meaning of are listed in Table 3.…”

Section: Table 2 Status Value Of Indicatorsmentioning

confidence: 99%

A Quantitative Scenario Analysis Method for Ecological Safety Development Trends in Rare Earth Mining Areas

Wang¹,

Liu²,

Ming-gui³

et al. 2019

JESTR

View full text Add to dashboard Cite

Rare earth mining can cause ecological problems as a result of soil erosion and water pollution in local areas. Thus assessing the development trends of ecological safety is crucial in formulating the environmental protection policies in rare earth area. Rare earth mining is a long and complicated process in which the state values of evaluation indicators for ecological safety are dynamic and continual interplay. Hence, conventionally measured indicators are inadequate in this context. An evaluation method based on the theory of quantitative scenario analysis was proposed in this study to analyze ecological safety development trends in rare earth mining areas. First, according to the PSR model framework, six indicators including "mining technology", "mining intensity", "water environment", "soil environment", "waste water, waste gas and waste residue management technologies" and "environmental protection policy", were selected to reveal the ecological safety in mining areas. Second, the crossover probability algorithm, Markov chain and nonlinear programming were utilized to construct a quantitative scenario analysis model for ecological safety development trends. Lastly, the model was verified using the data on Lingbei rare earth mining area located in Ganzhou City, China. Results showed that the quantitative scenario analysis model could be used to calculate the changes in various indicators, their cross impacts in the development process and the occurrence probability of scenario combinations for ecological safety development that were composed of the state changes of each indicator. These findings indicate that the proposed model can effectively and accurately forecast the ecological safety development trends in rare earth mining areas. The conclusions can provide a theoretical basis for environmental protection work in rare earth mining areas.

show abstract

A robust correlation analysis framework for imbalanced and dichotomous data with uncertainty

Abstract: This is a repository copy of A robust correlation analysis framework for imbalanced and dichotomous data with uncertainty.

Cited by 60 publications

References 45 publications

Industrial Data-Driven Monitoring Based on Incremental Learning Applied to the Detection of Novel Faults

Industrial Data-Driven Monitoring Based on Incremental Learning Applied to the Detection of Novel Faults

Load forecasting based on deep neural network and historical data augmentation

A Quantitative Scenario Analysis Method for Ecological Safety Development Trends in Rare Earth Mining Areas

Contact Info

Product

Resources

About