Performance of Stanford and Minipar Parser on Biomedical Texts

TL;DR

This paper evaluates the performance of Stanford and Minipar dependency parsers on biomedical texts, revealing limitations in their ability to accurately identify dependencies between connected biomedical concepts, especially over distant clauses.

Contribution

It provides a comparative analysis of Stanford and Minipar parsers on biomedical texts, highlighting their shortcomings and the impact of their underlying principles.

Findings

Minipar performs poorly compared to Stanford in biomedical parsing

Both parsers struggle with dependencies over distant clauses

Minipar's performance is inferior in precision, recall, and F-score

Abstract

In this paper, the performance of two dependency parsers, namely Stanford and Minipar, on biomedical texts has been reported. The performance of te parsers to assignm dependencies between two biomedical concepts that are already proved to be connected is not satisfying. Both Stanford and Minipar, being statistical parsers, fail to assign dependency relation between two connected concepts if they are distant by at least one clause. Minipar's performance, in terms of precision, recall and the F-score of the attachment score (e.g., correctly identified head in a dependency), to parse biomedical text is also measured taking the Stanford's as a gold standard. The results suggest that Minipar is not suitable yet to parse biomedical texts. In addition, a qualitative investigation reveals that the difference between working principles of the parsers also play a vital role for Minipar's degraded performance.