Uncertainty estimation and misclassification probability for classification models based on discriminant analysis and support vector machines

Morais, Camilo L. M.; Lima, Kássio M. G.; Martin, Francis

doi:10.1016/j.aca.2018.09.022

Cited by 33 publications

(21 citation statements)

References 24 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…1). While widely employed in the chemistry community,19,39,40,42,43 ensembles increase the model training effort in proportion to the number of models used (typically an order of magnitude, ESI Text S1†). Although this additional effort may be practical for some models ( e.g.…”

Section: Introductionmentioning

confidence: 99%

A quantitative uncertainty metric controls error in neural network-driven chemical discovery

et al. 2019

View full text Add to dashboard Cite

show abstract

Section: Introductionmentioning

confidence: 99%

A quantitative uncertainty metric controls error in neural network-driven chemical discovery

et al. 2019

View full text Add to dashboard Cite

show abstract

“…The kernel function (in this case, the radial bases function (RBF)) transforms the input spectral data into a feature space that maximises the margin of separation between the classes. Although more powerful than LDA or QDA for classification, SVM is more susceptible to overfitting 24 .…”

Section: Methodsmentioning

confidence: 99%

Spectrochemical differentiation in gestational diabetes mellitus based on attenuated total reflection Fourier-transform infrared (ATR-FTIR) spectroscopy and multivariate analysis

Bernardes-Oliveira

Freitas

Morais

et al. 2020

Sci Rep

Self Cite

View full text Add to dashboard Cite

Gestational diabetes mellitus (GDM) is a hyperglycaemic imbalance first recognized during pregnancy, and affects up to 22% of pregnancies worldwide, bringing negative maternal–fetal consequences in the short- and long-term. In order to better characterize GDM in pregnant women, 100 blood plasma samples (50 GDM and 50 healthy pregnant control group) were submitted Attenuated Total Reflection Fourier-transform infrared (ATR-FTIR) spectroscopy, using chemometric approaches, including feature selection algorithms associated with discriminant analysis, such as Linear Discriminant Analysis (LDA), Quadratic Discriminant Analysis (QDA) and Support Vector Machines (SVM), analyzed in the biofingerprint region between 1800 and 900 cm−1 followed by Savitzky–Golay smoothing, baseline correction and normalization to Amide-I band (~ 1650 cm−1). An initial exploratory analysis of the data by Principal Component Analysis (PCA) showed a separation tendency between the two groups, which were then classified by supervised algorithms. Overall, the results obtained by Genetic Algorithm Linear Discriminant Analysis (GA-LDA) were the most satisfactory, with an accuracy, sensitivity and specificity of 100%. The spectral features responsible for group differentiation were attributed mainly to the lipid/protein regions (1462–1747 cm−1). These findings demonstrate, for the first time, the potential of ATR-FTIR spectroscopy combined with multivariate analysis as a screening tool for fast and low-cost GDM detection.

show abstract

“…Moreover, each dimensionality reduction algorithm suggests a different set of variables to be used in SVMs, which can result in controversy. Recent studies ( [15], [18], and [24]) have shown that the boostrap procedure may have a number of benefits in classification problems, like reducing a model's uncertainty or finding the appropriate number of clusters in the k-nearest neighbor.…”

Section: E a R L Y B I R Dmentioning

confidence: 99%

Tenfold Bootstrap Procedure for Support Vector Machines

Vrigazova¹,

Ivanov²

2020

csci

View full text Add to dashboard Cite

Cross validation is often used to split input data into training and test set in Support vector machines. The two most commonly used cross validation versions are the tenfold and leave-one-out cross validation. Another commonly used resampling method is the random test/train split. The advantage of these methods is that they avoid overﬁtting in the model and perform model selection. They, however, can increase the computational time for ﬁtting Support vector machines with the increase of the size of the dataset. In this research, we propose an alternative for ﬁtting SVM, which we call the tenfold bootstrap for Support vector machines. This resampling procedure can signiﬁcantly reduce execution time despite the big number of observations, while preserving model’s accuracy. With this ﬁnding, we propose a solution to the problem of slow execution time when ﬁtting support vector machines on big datasets.

show abstract

Uncertainty estimation and misclassification probability for classification models based on discriminant analysis and support vector machines

Abstract: Uncertainty estimation and misclassification probability for classification models based on discriminant analysis and support vector machines,

Cited by 33 publications

References 24 publications

A quantitative uncertainty metric controls error in neural network-driven chemical discovery

A quantitative uncertainty metric controls error in neural network-driven chemical discovery

Spectrochemical differentiation in gestational diabetes mellitus based on attenuated total reflection Fourier-transform infrared (ATR-FTIR) spectroscopy and multivariate analysis

Tenfold Bootstrap Procedure for Support Vector Machines

Contact Info

Product

Resources

About