A EFFECTIVE APPROACH FOR SIMPLIFYING AGGREGATE MENTIONS IN BIOMEDICAL TEXT

Cloud Computing & Big Data

One of the major challenge in biomedical named entity recognition (NER) and normalization process is the detection and decision of aggregate(compound) named entities, in which a single entity refers to many concept e.g., SMAD/1/2. Previous research regarding named entity recognition and normalization, some of them have neglected aggregate mentions, apply simply rules for detecting, or perform coordination ellipsis, so that force to require a such method that can easily handle the different types of aggregate mentions. In this paper, we propose a new approach that combines a machine-learning approach with a pattern detection method to recognize the each entity of each aggregate mention. The proposed method effectively handles various types of aggregate mentions. The proposed method provides high performance in detecting and finding aggregate mentions that are: genes, diseases, and chemicals. The proposed system will later increase the performance of sequence as well as unwellness idea recognition, detection and normalization.

This work is licensed under a Creative Commons Attribution-ShareAlike 4.0 International License.