Time-Series Classification For Patients Under Active Surveillance And Screening Using Echo State Network
Li, Z.; Zaikin, A.; Blyuss, O.
Show abstract
AO_SCPLOWBSTRACTC_SCPLOWOver the past few decades, more and more patients come on follow-up studies such as active surveillance and screening, which results in a vast amount of time-series data in the health department. Each Patient typically has a small but different number of visits to the doctor and the time interval between the visits is heterogeneous. Nowadays, many machine learning tasks in relation to time series data are carried out using deep recurrent neural networks (RNN). However, deep neural networks consume enormous computational power as all weights in the network need to be trained through back-propagation. Conversely, echo state network (ESN), another form of RNN, demonstrates low training cost and the potential of it is still largely untapped. Therefore, in this article we will develop a new methodology that can classify aforementioned time-series data using the echo state network. We will also discuss how to address the heterogeneity in the time interval arising from the data of this type and how our model can also potentially fit other time-series data.
Matching journals
The top 12 journals account for 50% of the predicted probability mass.
Similar papers in this journal
- Data-driven Discovery of Mathematical and Physical Relations in Oncology Data using Human-understandable Machine Learning 97%
- PECLIDES Neuro - A Personalisable Clinical Decision Support System for Neurological Diseases 92%
- TDbasedUFE and TDbasedUFEadv: bioconductor packages to perform tensor decomposition based unsupervised feature extraction 91%
Similar papers in this journal
Similar papers in this journal
- ISMI-VAE: A Deep Learning Model for Classifying Disease Cells Using Gene Expression and SNV Data 95%
- Decoding Clinical Biomarker Space of COVID-19: Exploring Matrix Factorization-based Feature Selection Methods 95%
- BenchXAI: Comprehensive Benchmarking of Post-hoc Explainable AI Methods on Multi-Modal Biomedical Data 94%
"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.