TY - GEN
T1 - An active learning framework for efficient condition severity classification
AU - Nissim, Nir
AU - Boland, Mary Regina
AU - Moskovitch, Robert
AU - Tatonetti, Nicholas P.
AU - Elovici, Yuval
AU - Shahar, Yuval
AU - Hripcsak, George
N1 - Publisher Copyright:
© Springer International Publishing Switzerland 2015.
PY - 2015/1/1
Y1 - 2015/1/1
N2 - Understanding condition severity, as extracted from Electronic Health Records (EHRs), is important for many public health purposes. Methods requiring physicians to annotate condition severity are time-consuming and costly. Previously, a passive learning algorithm called CAESAR was developed to capture severity in EHRs. This approach required physicians to label conditions manually, an exhaustive process. We developed a framework that uses two Active Learning (AL) methods (Exploitation and Combination_XA) to decrease manual labeling efforts by selecting only the most informative conditions for training. We call our approach CAESAR-Active Learning Enhancement (CAESAR-ALE). As compared to passive methods, CAESAR-ALE’s first AL method, Exploitation, reduced labeling efforts by 64% and achieved an equivalent true positive rate, while CAESARALE’s second AL method, Combination_XA, reduced labeling efforts by 48% and achieved equivalent accuracy. In addition, both these AL methods outperformed the traditional AL method (SVM-Margin). These results demonstrate the potential of AL methods for decreasing the labeling efforts of medical experts, while achieving greater accuracy and lower costs.
AB - Understanding condition severity, as extracted from Electronic Health Records (EHRs), is important for many public health purposes. Methods requiring physicians to annotate condition severity are time-consuming and costly. Previously, a passive learning algorithm called CAESAR was developed to capture severity in EHRs. This approach required physicians to label conditions manually, an exhaustive process. We developed a framework that uses two Active Learning (AL) methods (Exploitation and Combination_XA) to decrease manual labeling efforts by selecting only the most informative conditions for training. We call our approach CAESAR-Active Learning Enhancement (CAESAR-ALE). As compared to passive methods, CAESAR-ALE’s first AL method, Exploitation, reduced labeling efforts by 64% and achieved an equivalent true positive rate, while CAESARALE’s second AL method, Combination_XA, reduced labeling efforts by 48% and achieved equivalent accuracy. In addition, both these AL methods outperformed the traditional AL method (SVM-Margin). These results demonstrate the potential of AL methods for decreasing the labeling efforts of medical experts, while achieving greater accuracy and lower costs.
KW - Active-learning
KW - Condition
KW - Electronic health records
KW - Phenotyping
UR - http://www.scopus.com/inward/record.url?scp=84947910777&partnerID=8YFLogxK
U2 - 10.1007/978-3-319-19551-3_3
DO - 10.1007/978-3-319-19551-3_3
M3 - Conference contribution
AN - SCOPUS:84947910777
SN - 9783319195506
T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
SP - 13
EP - 24
BT - Artificial Intelligence in Medicine - 15th Conference on Artificial Intelligence in Medicine, AIME 2015, Proceedings
A2 - Bellazzi, Riccardo
A2 - Sacchi, Lucia
A2 - Holmes, John H.
A2 - Peek, Niels
PB - Springer Verlag
T2 - 15th Conference on Artificial Intelligence in Medicine, AIME 2015
Y2 - 17 June 2015 through 20 June 2015
ER -