Sequential Labeling With Structural SVM Under Nondecomposable Losses.

Sequential labeling addresses the classification of sequential data, which are widespread in fields as diverse as computer vision, finance, and genomics. The model traditionally used for sequential labeling is the hidden Markov model (HMM), where the sequence of class labels to be predicted is encoded as a Markov chain. In recent years, HMMs have benefited from minimum-loss training approaches, such as the structural support vector machine (SSVM), which, in many cases, has reported higher classification accuracy.

Interpolation to define clinical tumor stage in prostate cancer using clinical description of digital rectal examination.

Aim: This study aims to assess characteristics of patients with prostate cancer for whom clinical T stage category (cT) was not documented in the medical record and assess whether specialists had concordant conclusions regarding cT based on digital rectal examination (DRE) notes.

Methods: Data from the Prostate Cancer Outcome Registry - Victoria (PCOR-Vic) were interrogated. Four specialists independently assigned cT to DRE notes.

Recurrent neural networks with specialized word embeddings for health-domain named-entity recognition.

Background: Previous state-of-the-art systems on Drug Name Recognition (DNR) and Clinical Concept Extraction (CCE) have focused on a combination of text "feature engineering" and conventional machine learning algorithms such as conditional random fields and support vector machines. However, developing good features is inherently heavily time-consuming. Conversely, more modern machine learning approaches such as recurrent neural networks (RNNs) have proved capable of automatically learning effective features from either random assignments or automated word "embeddings".