Efficient leave-one-out cross-validation of kernel fisher discriminant classifiers

Cawley, Gavin C.; Talbot, Nicola L. C.

doi:10.1016/s0031-3203(03)00136-5

Cited by 388 publications

(245 citation statements)

References 15 publications

Supporting

Mentioning

243

Contrasting

Unclassified

Order By: Relevance

“…Next, PLSR was conducted with centered and scaled variables using leave-one-out cross-validation (LOOCV). LOOCV is effective on small data sets and has low bias, variance, and error rates compared to other cross-validation techniques (Cawley and Talbot 2003;Mevik and Cederkvist 2004). For each nutrient, the PLSR model with the minimum root mean PRESS was selected.…”

Section: Indexmentioning

confidence: 99%

Predicting macronutrient concentrations from loblolly pine leaf reflectance across local and regional scales

Stein

Thomas

Lorentz

et al. 2014

GIScience & Remote Sensing

View full text Add to dashboard Cite

Given the economic importance of loblolly pine (Pinus taeda) in the southeastern US, there is a need to establish efficient methods of detecting potential nutrient deficiencies that may limit productivity. This study evaluated the use of remote sensing for macronutrient assessment in loblolly pine. Reflectance-based models were developed at two spatial scales: (1) a natural nutrient gradient across the species' range, and (2) localized fertilization and genotype treatments in North Carolina and Virginia. Fascicles were collected regionally from 237 samples of 3 flushes at 18 sites, and locally from 72 trees with 2 fertilization treatments and 6 genotypes. Sample spectral reflectance was calculated using a spectroradiometer, and nutrient concentrations were measured with dry combustion and wet chemical digestion. Results were analyzed statistically using nutrient correlations with reflectance and common vegetation indices, and partial least squares regression (PLSR). PLSR performed well at the regional scale, with R 2 values for nitrogen, phosphorus, potassium, calcium, and magnesium of 0. 81, 0.70, 0.68, 0.42, and 0.51, respectively. No model successfully predicted nutrients at local sites for any treatment or canopy stratum. This discrepancy implies that a large nutrient range and/or spatial scale may be necessary to model loblolly pine nutrients with spectral reflectance.

show abstract

Section: Indexmentioning

confidence: 99%

Predicting macronutrient concentrations from loblolly pine leaf reflectance across local and regional scales

Stein

Thomas

Lorentz

et al. 2014

GIScience & Remote Sensing

View full text Add to dashboard Cite

show abstract

“…Leave-many-out (LMO) is an unbiased technique for crossvalidation of classification results [19,20]. We used this technique to test our classifier.…”

Section: Classification Resultsmentioning

confidence: 99%

Disease classification: A probabilistic approach

Rathi

Malcolm

Bouix

et al. 2010

2010 IEEE International Symposium on Biomedical Imaging: From Nano to Macro

View full text Add to dashboard Cite

We describe a probabilistic technique for separating two populations whereby analysis is performed on affine-invariant representations of each patient. The method begins by converting each voxel from a high-dimensional diffusion weighted signal to a low-dimensional diffusion tensor representation. Three orthogonal measures that capture different aspects of the local tissue are derived from the tensor representation to form a feature vector. From these feature vectors, we form a probabilistic representation of each patient. This representation is affine invariant, thus obviating the need for registration of the images. We then use a Parzen window classifier to estimate the likelihood of a new patient belonging to either population. To demonstrate the technique, we apply it to the analysis of 22 first-episode schizophrenic patients and 20 normal control subjects. With leave-many-out cross validation, we find a detection rate of 90.91% (10% false positives).

show abstract

“…We then assess the cause of high MAE by categorizing stress level into two classes -"1" for non-stressed, and "2" for stressed (see Table IV). Classifiers were then trained using the scaled target class, and cross-validated using leave-one-out cross-validation (LOOCV) [35]. The corresponding results of all test sets are presented in Table V. KNN is able to achieve 72% accuracy, which performs much better than ANN (44%) and LDA (60%).…”

Section: B Data Classificationmentioning

confidence: 99%

Analysis of Single-Electrode EEG Rhythms Using MATLAB to Elicit Correlation with Cognitive Stress

Lim¹,

Chia²

2015

IJCTE

View full text Add to dashboard Cite

Abstract-This paper demonstrates electroencephalogram (EEG) analysis in MATLAB environment with the objective to investigate effectiveness of cognitive stress recognition algorithm using EEG from single-electrode BCI. 25 subjects' EEG were recorded in MATLAB with the use of Stroop color-word test as stress inducer. Questionnaire on subjects' self-perceived stress scale during Stroop test were gathered as classification's target output. The main analysis tool used were MATLAB, coupled with the use of Discrete Cosine Transform (DCT) as dimension reduction technique to reduce data size down to 2% of the origin. Three pattern classification algorithms' -Artificial Neural Network (ANN), k-Nearest Neighbor (KNN) and Linear Discriminant Analysis (LDA) were trained using the resulted 2% DCT coefficients. Our study discovered the use of DCT along with KNN offers highest average classification rate of 72% compared to ANN and LDA.

show abstract

Efficient leave-one-out cross-validation of kernel fisher discriminant classifiers

Cited by 388 publications

References 15 publications

Predicting macronutrient concentrations from loblolly pine leaf reflectance across local and regional scales

Predicting macronutrient concentrations from loblolly pine leaf reflectance across local and regional scales

Disease classification: A probabilistic approach

Analysis of Single-Electrode EEG Rhythms Using MATLAB to Elicit Correlation with Cognitive Stress

Contact Info

Product

Resources

About