A hybrid deep learning approach for medical relation extraction
This work addresses a domain-specific need for medical relation extraction to support applications like decision support systems, but it is incremental as it combines existing methods.
The paper tackled the problem of extracting relationships between treatments and medical problems in biomedical texts by proposing a hybrid deep learning and rule-based approach, achieving promising performance on the I2b2 2010 dataset.
Mining relationships between treatment(s) and medical problem(s) is vital in the biomedical domain. This helps in various applications, such as decision support system, safety surveillance, and new treatment discovery. We propose a deep learning approach that utilizes both word level and sentence-level representations to extract the relationships between treatment and problem. While deep learning techniques demand a large amount of data for training, we make use of a rule-based system particularly for relationship classes with fewer samples. Our final relations are derived by jointly combining the results from deep learning and rule-based models. Our system achieved a promising performance on the relationship classes of I2b2 2010 relation extraction task.