The generalisability of artificial neural networks used to classify electrophoretic data produced under different conditions

Taylor, Duncan; Kitselaar, Michael; Powers, David

doi:10.1016/j.fsigen.2018.10.019

Cited by 11 publications

(8 citation statements)

References 3 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Bookmaker informedness has been used in several studies, most authored by Powers [42][43][44], and few authored by other scientists (two in image recognition [45,46], one in text mining [47], and one in vehicle tracking [48]). Two studies show the effectiveness of markedness in environmental sciences [49] and economics [50].…”

Section: Introductionmentioning

confidence: 99%

The Matthews correlation coefficient (MCC) is more reliable than balanced accuracy, bookmaker informedness, and markedness in two-class confusion matrix evaluation

2021

View full text Add to dashboard Cite

Evaluating binary classifications is a pivotal task in statistics and machine learning, because it can influence decisions in multiple areas, including for example prognosis or therapies of patients in critical conditions. The scientific community has not agreed on a general-purpose statistical indicator for evaluating two-class confusion matrices (having true positives, true negatives, false positives, and false negatives) yet, even if advantages of the Matthews correlation coefficient (MCC) over accuracy and F1 score have already been shown.In this manuscript, we reaffirm that MCC is a robust metric that summarizes the classifier performance in a single value, if positive and negative cases are of equal importance. We compare MCC to other metrics which value positive and negative cases equally: balanced accuracy (BA), bookmaker informedness (BM), and markedness (MK). We explain the mathematical relationships between MCC and these indicators, then show some use cases and a bioinformatics scenario where these metrics disagree and where MCC generates a more informative response.Additionally, we describe three exceptions where BM can be more appropriate: analyzing classifications where dataset prevalence is unrepresentative, comparing classifiers on different datasets, and assessing the random guessing level of a classifier. Except in these cases, we believe that MCC is the most informative among the single metrics discussed, and suggest it as standard measure for scientists of all fields. A Matthews correlation coefficient close to +1, in fact, means having high values for all the other confusion matrix metrics. The same cannot be said for balanced accuracy, markedness, bookmaker informedness, accuracy and F1 score.

show abstract

Section: Introductionmentioning

confidence: 99%

The Matthews correlation coefficient (MCC) is more reliable than balanced accuracy, bookmaker informedness, and markedness in two-class confusion matrix evaluation

2021

View full text Add to dashboard Cite

show abstract

“…The normalization and baselining of the raw data are carried out so that training data can apply to test data from either platform. Training data for the current work includes all training data from [5] as well as additional profiles to assist with classifying low level data. In total the input dataset comprised 703 profiles from:…”

Section: Ann Input Datamentioning

confidence: 99%

“…To demonstrate this fact, and investigate the performance of the MHCNN within a profile reading framework, the MHCNN was embedded in the DNA profile reading software FaSTR™ DNA. A run of reference profiles containing 48 profiles (1700 allelic peaks) was chosen that exhibited a number of misclassifications when using the original NN from [5]. Table 4 shows the comparison of performance between the original NN and the MHCNN.…”

Section: Test Setmentioning

confidence: 99%

See 1 more Smart Citation

Using a multi-head, convolutional neural network with data augmentation to improve electropherogram classification performance

Taylor

2022

Forensic Science International: Genetics

Self Cite

View full text Add to dashboard Cite

“…Deep neural networks were also applied in predicting geographic location using whole-genome sequence data of the organisms, achieving median test errors of 16.9km, 5.7km and 85km for three species (Plasmodium parasites, Anopheles mosquitoes, and global human populations) (16). More specifically, artificial neural networks were also used in classifying electrophoresis profiles in forensic casework (17,18). Here in this study, we used machine learning to predict Y haplogroups to a fine resolution based on Y-STRs.…”

Section: Introductionmentioning

confidence: 99%

YHP: Y-chromosome Haplogroup Predictor for predicting male lineages based on Y-STRs

Song

Zhao

et al. 2021

Preprint

View full text Add to dashboard Cite

Human Y chromosome reflects the evolutionary process of males. Male lineage tracing by Y chromosome is of great use in evolutionary, forensic, and anthropological studies when male samples exist or especially when the biological sample is a mixture of male and female individuals. Identifying the male lineage based on the specific distribution of Y haplogroups narrows down the investigation scope. Integrating previously published datasets with genotypes of Y chromosome short tandem repeats (Y-STRs) and high-resolution haplogroups (122 haplogroups in total), we developed YHP (Y Haplogroup Predictor), an open-access and userfriendly software package to predict haplogroups, compare the similarity, and conduct mismatch analysis of samples with Y-STR profiles. The software is available at Github (https://github.com/cissy123/YHP-Y-Haplogroup-Predictor-).Author SummaryFamilial searching has been used in forensic, anthropologic, and personalized scenarios. Software packages have been developed to assist in male familial searching, such as predicting Y-SNP haplogroups by Y-STRs. However, these software packages, in general, achieve this goal with a rough resolution. In this study, we developed a software package to conduct high-resolution haplogroup inference to help familial searching and at the same time reduce the cost, since it does not require tiresome Y-SNP sequencing.

show abstract

The generalisability of artificial neural networks used to classify electrophoretic data produced under different conditions

Cited by 11 publications

References 3 publications

The Matthews correlation coefficient (MCC) is more reliable than balanced accuracy, bookmaker informedness, and markedness in two-class confusion matrix evaluation

The Matthews correlation coefficient (MCC) is more reliable than balanced accuracy, bookmaker informedness, and markedness in two-class confusion matrix evaluation

Using a multi-head, convolutional neural network with data augmentation to improve electropherogram classification performance

YHP: Y-chromosome Haplogroup Predictor for predicting male lineages based on Y-STRs

Contact Info

Product

Resources

About