Biomedical relation extraction using stochastic difference equations

We propose an unsupervised method for extracting causal relations between biomedical entities using stochastic difference equations (SDE). Our method attempts to generalize the propagation of relevance in medical sentences in order to extract related biomedical terms and the semantic relation between them. We model the propagation of relevance in candidate medical sentences through the use of stochastic difference equations. The equation of propagation of relevance helps in identifying the most relevant medical terms in a causal sentence. It also increases the accuracy of Information Extraction as it allows to set a threshold for the minimum relevance for relation extraction.