Assessing the accuracy of prediction algorithms for classification: an overview

By Pierre Baldi, Søren Brunak, Yves Chauvin, Claus A. F. Andersen, Henrik Nielsen. Bioinform., 2000

2,199 citations recorded by the source corpus.

View Paper: https://www.semanticscholar.org/paper/78200101cb94b926b376f38571bffcaf0f9cc50e

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Molecular Biology

Citations in the corpus

A Systematic Prediction of Multiple Drug-Target Interactions from Chemical, Genomic, and Pharmacological Data (2012, background): “In the case of classification, the assessment of the prediction quality of statistical models is typically performed on the basis of several parameters [35].”
An automated method for finding molecular complexes in large protein interaction networks (2003, neutral): “A specificity versus sensitivity analysis [32] was also performed.”
Revealing Household Characteristics from Smart Meter Data (2014, background): “, total disagreement between the ground truth data and the estimation) and 1 (total agreement), whereas the MCC is 0 for random estimations [46].”
AlgPred: prediction of allergenic proteins and mapping of IgE epitopes (2006, background): “Following is a brief description of the parameters (32): (i) sensitivity, also referred to as recall, is the percent of correctly predicted allergen epitopes (15); (ii) specificity is the percent of correctly predicted non-allergen epitopes; (iii) accuracy is the proportion of correctly predicted ep”
Prediction of protein subcellular localization (2006, methodology): “We use the standard Qi percentage accuracy[36-38] and the Matthew's correlation coefficient[39] MCC for assessing the accuracy of subcellular”
Predicting protein-protein interactions from primary structure (2001, background): “There is a negligible probability that the learning system will ‘learn its own input’ (see Baldi et al., 2000 ) on a narrow, highly self-similar set of data examples.”
Predicting linear B-cell epitopes using string kernels (2008, neutral): “The prediction accuracy (ACC), sensitivity ( S n ), specificity ( S p ), and correlation coefficient ( CC ) are often used to evaluate prediction algorithms (Baldi et al ., 2000).”
EGASP: the human ENCODE Genome Annotation Assessment Project (2006, neutral): “[24], Guigó et al .”
Performance of mutation pathogenicity prediction methods on missense variants (2011, result): “Because of its insensitivity to differing test set sizes, it gives a more balanced assessment of performance than the other performance measures [Baldi et al., 2000].”
ClaRNA: a classifier of contacts in RNA 3D structures based on a comparative analysis of various classification schemes (2014, methodology): “We also evaluated the parameters using the methodology from (21) and obtained an overall percentage of correct predictions Q total = 99.”
Learning to Detect Phishing Webpages (2014, background): “MCC takes into account true and false positive and negative and is generally regarded as a balanced measure that can be used even if the classes are of very different sizes [10].”
Improving the prediction of protein secondary structure in three and eight classes using recurrent neural networks and profiles (2002, neutral): “Several indices can be used to score the e ciency of the algorithm [5].”
A statistical framework for neuroimaging data analysis based on mutual information estimated via a gaussian copula (2016, result): “Similar in spirit to Matthews’ correlation coefficient [Baldi et al., 2000; Matthews, 1975], this provides an overall accuracy measure, which incorporates possible asymmetries resulting from unbalanced classes.”
ROC analysis: applications to the classification of biological sequences and 3D structures (2008, methodology): “Many similar indices are reviewed in [2] and [3].”