Patient-specific early classification of multivariate observations

Abstract

Early classification of time series has been receiving a lot of attention recently. In this paper we present a model, which we call the Early Classification Model (ECM), that allows for early, accurate and patient-specific classification of multivariate observations. ECM is comprised of an integration of the widely used Hidden Markov Model (HMM) and Support Vector Machine (SVM) models. It attained very promising results on the datasets we tested it on: in one set of experiments based on a published dataset of response to drug therapy in Multiple Sclerosis patients, ECM used only an average of 40% of a time series and was able to outperform some of the baseline models, which needed the full time series for classification. In the set of experiments tested on a sepsis therapy dataset, ECM was able to surpass the standard threshold-based method and the state-of-the-art method for early classification of multivariate time series.

Inferior Colliculus Transcriptome After Status Epilepticus in the Genetically Audiogenic Seizure-Prone Hamster GASH/Sal

“…In addition, an overrepresentation analysis of the total data was performed to identify any priority pathway in the epileptogenic focus. To determine the gene expression level groups, k-means clustering was performed (MacQueen, 1967). This clustering makes it possible to establish relationships between distant genes based on the assumption that related genes will be expressed similarly.…”

“…Of the top differentially expressed genes, 21 of them have a human homolog described in the metabolome. These 21 genes were clustered according to their gene expression levels by k-means clustering (MacQueen, 1967), which resulted in 3 clusters (Figure 7). In cluster 1, the Rxfp2 gene stands out from the other study genes as a single-member cluster (red circle, Figure 7).…”

Extraction of Interpretable Multivariate Patterns for Early Diagnostics

“…A method called hybrid HMM/SVM was proposed for early classification of multivariate time series [28], [29]. The method uses HMMs to examine all segments from the original time series and generate likelihood of the membership of the pattern which is then passed to SVM to decided the probability of the class membership of the time series.…”

Artificial neural network classification of microarray data using new hybrid gene selection method

“…Many of the researchers have employed multi-step dimensionality reduction methods for feature selection and semi-supervised learning methods for classification of cancer (Gui et al, 2010;Hernández-Lobato et al, 2010). Various intelligent machine learning techniques have been applied to this area, such as Support Vector Machine (SVM), Naïve Bayes (NB), Decision Tree, Random Forest and kernel-based classifiers, have been successfully applied to microarray data classification in the recent years (Ghalwash et al, 2015;Mohan et al, 2014;Peng, 2006). In contrast, methods such as Artificial Neural Networks (ANNs) are able to recognise subtly different biological entities.…”