Semi-Supervised Learning to Identify UMLS Semantic Relations

Dynamic categorization of clinical research eligibility criteria by hierarchical clustering

OBJECTIVE To semi-automatically induce semantic categories of eligibility criteria from text and to automatically classify eligibility criteria based on their semantic similarity. DESIGN The UMLS semantic types and a set of previously developed semantic preference rules were utilized to create an unambiguous semantic feature representation to induce eligibility criteria categories through hie...

Clustering UMLS Semantic Relations Between Medical Concepts

We propose and implement an innovative semi-supervised framework for automatically discovering UMLS semantic relations. Our proposed framework uses semantic, syntactic and ortho-graphic features both at global level and local level. We experimented with multiple distance metric for clustering including Euclidean distance, spherical k-means distance, and Kullback-Leibler divergence. We show that...

From efficiency to portability: acquisition of semantic relations by semi-supervised machine learning

Semi-Supervised Convolution Graph Kernels for Relation Extraction

Extracting semantic relations between entities is an important step towards automatic text understanding. In this paper, we propose a novel Semi-supervised Convolution Graph Kernel (SCGK) method for semantic Relation Extraction (RE) from natural English text. By encoding sentences as dependency graphs of words, SCGK computes kernels (similarities) between sentences using a convolution strategy,...

Automatic Identification of Treatment Relations For Medical Ontology Learning: An Exploratory Study

This study is part of a project to develop an automatic method to build ontologies, especially in a medical domain, from a document collection. An earlier study had investigated an approach to inferring semantic relations between medical concepts using the UMLS (Unified Medical Language System) semantic net. The study found that semantic relations between concepts could be inferred 68% of the t...