↓ Skip to main content

An active learning based classification strategy for the minority class problem: application to histopathology annotation

Overview of attention for article published in BMC Bioinformatics, October 2011
Altmetric Badge

Mentioned by

twitter
1 tweeter

Citations

dimensions_citation
53 Dimensions

Readers on

mendeley
95 Mendeley
citeulike
3 CiteULike
You are seeing a free-to-access but limited selection of the activity Altmetric has collected about this research output. Click here to find out more.
Title
An active learning based classification strategy for the minority class problem: application to histopathology annotation
Published in
BMC Bioinformatics, October 2011
DOI 10.1186/1471-2105-12-424
Pubmed ID
Authors

Scott Doyle, James Monaco, Michael Feldman, John Tomaszewski, Anant Madabhushi

Abstract

Supervised classifiers for digital pathology can improve the ability of physicians to detect and diagnose diseases such as cancer. Generating training data for classifiers is problematic, since only domain experts (e.g. pathologists) can correctly label ground truth data. Additionally, digital pathology datasets suffer from the "minority class problem", an issue where the number of exemplars from the non-target class outnumber target class exemplars which can bias the classifier and reduce accuracy. In this paper, we develop a training strategy combining active learning (AL) with class-balancing. AL identifies unlabeled samples that are "informative" (i.e. likely to increase classifier performance) for annotation, avoiding non-informative samples. This yields high accuracy with a smaller training set size compared with random learning (RL). Previous AL methods have not explicitly accounted for the minority class problem in biomedical images. Pre-specifying a target class ratio mitigates the problem of training bias. Finally, we develop a mathematical model to predict the number of annotations (cost) required to achieve balanced training classes. In addition to predicting training cost, the model reveals the theoretical properties of AL in the context of the minority class problem.

Twitter Demographics

The data shown below were collected from the profile of 1 tweeter who shared this research output. Click here to find out more about how the information was compiled.

Mendeley readers

The data shown below were compiled from readership statistics for 95 Mendeley readers of this research output. Click here to see the associated Mendeley record.

Geographical breakdown

Country Count As %
Brazil 4 4%
Colombia 2 2%
Germany 1 1%
Netherlands 1 1%
France 1 1%
Sweden 1 1%
United Kingdom 1 1%
United States 1 1%
Unknown 83 87%

Demographic breakdown

Readers by professional status Count As %
Student > Ph. D. Student 19 20%
Researcher 17 18%
Student > Bachelor 10 11%
Student > Master 9 9%
Other 7 7%
Other 23 24%
Unknown 10 11%
Readers by discipline Count As %
Computer Science 31 33%
Medicine and Dentistry 15 16%
Engineering 12 13%
Agricultural and Biological Sciences 9 9%
Chemistry 6 6%
Other 11 12%
Unknown 11 12%

Attention Score in Context

This research output has an Altmetric Attention Score of 1. This is our high-level measure of the quality and quantity of online attention that it has received. This Attention Score, as well as the ranking and number of research outputs shown below, was calculated when the research output was last mentioned on 31 October 2011.
All research outputs
#7,762,553
of 12,373,386 outputs
Outputs from BMC Bioinformatics
#3,177
of 4,576 outputs
Outputs of similar age
#61,503
of 104,421 outputs
Outputs of similar age from BMC Bioinformatics
#13
of 16 outputs
Altmetric has tracked 12,373,386 research outputs across all sources so far. This one is in the 23rd percentile – i.e., 23% of other outputs scored the same or lower than it.
So far Altmetric has tracked 4,576 research outputs from this source. They receive a mean Attention Score of 4.9. This one is in the 21st percentile – i.e., 21% of its peers scored the same or lower than it.
Older research outputs will score higher simply because they've had more time to accumulate mentions. To account for age we can compare this Altmetric Attention Score to the 104,421 tracked outputs that were published within six weeks on either side of this one in any source. This one is in the 25th percentile – i.e., 25% of its contemporaries scored the same or lower than it.
We're also able to compare this research output to 16 others from the same source and published within six weeks on either side of this one. This one is in the 12th percentile – i.e., 12% of its contemporaries scored the same or lower than it.