Calculation of the reliability of classification in discriminant partial least-squares binary classification

Pérez, Néstor F.; Ferré, Joan; Boqué, Ricard

doi:10.1016/j.chemolab.2008.09.005

Cited by 144 publications

(78 citation statements)

References 20 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The model predicted twoŷ-values and Σŷ i =1. An often used approach for assigning the membership of a class is the "winner-takes-all-strategy" and the majority vote [30]. This means that the highest score calculated from the model obtains the class-assignment.…”

Section: Methodsmentioning

confidence: 99%

Evaluating Flood Exposure for Properties in Urban Areas Using a Multivariate Modelling Technique

Torgersen

Rød

Kvaal

et al. 2017

Water

View full text Add to dashboard Cite

Urban flooding caused by heavy rainfall is expected to increase in the future. The main purpose of this study was to investigate the variables characterizing the placement of a house, which seem to have an impact when it comes to the exposure to floods. From the same region in Norway, data from 347 addresses were derived. All addresses were either associated with insurance claims caused by flooding or were randomly selected. A multivariate statistical model, Partial Least Square Regression (PLS), was used. Among others, the analysis has shown that the upstream, sealed area is the most significant variable for characterizing properties' exposure to urban flooding. The model confirms that flooding tends to occur near old combined sewer mains and in concave curvature, and houses located in steep slopes seem to be less exposed. Using this method, it is possible to rank and quantify significant exposure variables contributing to urban floods within a region. Results from the PLS-analysis might provide important input to professionals, when planning and prioritizing measures. It can also predict flood-prone areas and make residents aware of the risks, which may induce them to implement preventive measures.

show abstract

Section: Methodsmentioning

confidence: 99%

Evaluating Flood Exposure for Properties in Urban Areas Using a Multivariate Modelling Technique

Torgersen

Rød

Kvaal

et al. 2017

Water

View full text Add to dashboard Cite

show abstract

“…Though ELM tends to provide better generalization performance at a fast learning speed and relative simplicity of use [12], ELM algorithm may have uncertainty in different trials of prediction due to the stochastic initialization of input weights and bias, which would make the classification of the raw data unreliable. Pérez et al (2009) proposed probabilistic discriminant partial least squares (p-DPLS) to improve the reliability of the classification by integrating density methods and Bayes decision theory [13]. Zhao et al (2011) proposed a binary probabilistic extreme learning machine (PELM) classification method to enhance the reliability of classification and avoid the misclassification due to the uncertainty of ELM predictions [14].…”

Section: Introductionmentioning

confidence: 99%

Multi-class Classification with One-Against-One Using Probabilistic Extreme Learning Machine

Zhao

Chai

Diao

et al. 2012

Advances in Neural Networks – ISNN 2012

View full text Add to dashboard Cite

Abstract. Probabilistic extreme learning machine (PELM) is a binary classification method, which can improve the computational speed, generalization performance and computational cost. In this work we extend the binary PELM to resolve multi-class classification problems by using oneagainst-one (OAO) and winner-takes-all strategy. The strategy one-against-one (OAO) involves C(C-1)/2 binary PELM models. A reliability for each sample is calculated from each binary PELM model, and the sample is assigned to the class with the largest combined reliability by using the winner-takes-all strategy. The proposed method is verified with the operational conditions classification of an industrial wastewater treatment plant. Experimental results show the good performance on classification accuracy and computational expense.

show abstract

“…The classification of samples is conducted by choosing a label with the highest probability, and usually a threshold is used to determine which class a sample belongs to. In two-class scenarios with labels "0" and "1", we assume that such a threshold equals 0.5 for simplicity, although the Bayes theorem can be used to give a more rigorous result [38].…”

Section: Partial Least Squares Discriminant Analysis (Pls-da)mentioning

confidence: 99%

Outlier Detection in Dynamic Systems with Multiple Operating Points and Application to Improve Industrial Flare Monitoring

Bell

et al. 2017

Processes

View full text Add to dashboard Cite

Abstract:In chemical industries, process operations are usually comprised of several discrete operating regions with distributions that drift over time. These complexities complicate outlier detection in the presence of intrinsic process dynamics. In this article, we consider the problem of detecting univariate outliers in dynamic systems with multiple operating points. A novel method combining the time series Kalman filter (TSKF) with the pruned exact linear time (PELT) approach to detect outliers is proposed. The proposed method outperformed benchmark methods in outlier removal performance using simulated data sets of dynamic systems with mean shifts.

show abstract

Calculation of the reliability of classification in discriminant partial least-squares binary classification

Cited by 144 publications

References 20 publications

Evaluating Flood Exposure for Properties in Urban Areas Using a Multivariate Modelling Technique

Evaluating Flood Exposure for Properties in Urban Areas Using a Multivariate Modelling Technique

Multi-class Classification with One-Against-One Using Probabilistic Extreme Learning Machine

Outlier Detection in Dynamic Systems with Multiple Operating Points and Application to Improve Industrial Flare Monitoring

Contact Info

Product

Resources

About