This Article 
   
 Share 
   
 Bibliographic References 
   
 Add to: 
 
Digg
Furl
Spurl
Blink
Simpy
Google
Del.icio.us
Y!MyWeb
 
 Search 
   
A Framework for Mining Signatures from Event Sequences and Its Applications in Healthcare Data
Feb. 2013 (vol. 35 no. 2)
pp. 272-285
Fei Wang, IBM T.J. Watson Res. Center, Hawthorne, NY, USA
Noah Lee, Dept. of Biomed. Eng., Columbia Univ., New York, NY, USA
Jianying Hu, IBM T.J. Watson Res. Center, Hawthorne, NY, USA
Jimeng Sun, IBM T.J. Watson Res. Center, Hawthorne, NY, USA
S. Ebadollahi, IBM T.J. Watson Res. Center, Hawthorne, NY, USA
A. F. Laine, Dept. of Biomed. Eng., Columbia Univ., New York, NY, USA
This paper proposes a novel temporal knowledge representation and learning framework to perform large-scale temporal signature mining of longitudinal heterogeneous event data. The framework enables the representation, extraction, and mining of high-order latent event structure and relationships within single and multiple event sequences. The proposed knowledge representation maps the heterogeneous event sequences to a geometric image by encoding events as a structured spatial-temporal shape process. We present a doubly constrained convolutional sparse coding framework that learns interpretable and shift-invariant latent temporal event signatures. We show how to cope with the sparsity in the data as well as in the latent factor model by inducing a double sparsity constraint on the β-divergence to learn an overcomplete sparse latent factor model. A novel stochastic optimization scheme performs large-scale incremental learning of group-specific temporal event signatures. We validate the framework on synthetic data and on an electronic health record dataset.
Index Terms:
stochastic programming,data mining,health care,knowledge representation,learning (artificial intelligence),medical information systems,electronic health record dataset,heterogeneous event sequences,healthcare data,temporal knowledge representation,learning framework,large-scale temporal signature mining,longitudinal heterogeneous event data,high-order latent event structure representation,high-order latent event structure extraction,high-order latent event structure mining,geometric image,event encoding,structured spatial-temporal shape process,doubly constrained convolutional sparse coding framework,interpretable latent temporal event signature learning,shift-invariant latent temporal event signatures,double sparsity constraint,β-divergence,overcomplete sparse latent factor model,stochastic optimization scheme,large-scale incremental learning,group-specific temporal event signatures,synthetic data,Convolution,Sparse matrices,Knowledge representation,Data mining,Complexity theory,Approximation methods,Convergence,beta-divergence,Temporal signature mining,sparse coding,dictionary learning,nonnegative matrix factorization,stochastic gradient descent
Citation:
Fei Wang, Noah Lee, Jianying Hu, Jimeng Sun, S. Ebadollahi, A. F. Laine, "A Framework for Mining Signatures from Event Sequences and Its Applications in Healthcare Data," IEEE Transactions on Pattern Analysis and Machine Intelligence, vol. 35, no. 2, pp. 272-285, Feb. 2013, doi:10.1109/TPAMI.2012.111
Usage of this product signifies your acceptance of the Terms of Use.