Molecules are a fundamental building block of life, critical to many aspects of scientific and technical work. Informatics related to molecules, compounds, and materials, their properties, structure, and behavior, can provide a significant acceleration to scientific and technical innovation. A first step in this process is to extract relevant information from unstructured data such as patents and journal articles. Current tools focus on the text mining of small molecule names using either dictionary and rule-based models or statistical machine learning methods. Extracting information about polymers and their synthetic parameters is rare, and so far either inaccurate or incomplete. In this talk, we will introduce a polymer annotator based on IBM’s SystemT, which adopts an algebraic approach.  We will show examples of extracted data, compare performance to manual data extraction and present how the annotator is incorporated into the cheminformatics tool CIRCA.  <br/>

Extraction of polymer-related information in the cheminformatics tool CIRCA

2019 Fall National Meeting

The official archive and repository for all ACS Meetings & Expositions abstracts and select posters or presentations.