A Performance Analysis Framework for Coreference Resolution Algorithms

This thesis entitled A Performance Analysis Framework for Coreference Resolution Algorithms, focuses on the topic of coreference resolution of semantic datasets. In order for Big Data analytics to be effective, it is essential to develop automated algorithms capable of integrating multiple datasets that contain data about a particular person or other entity. Accomplishing this necessitates coreference resolution; for example, determining that J. Doe in one dataset refers to the same person as Jonathan Doe Jr. in another dataset. There are many existing coreference resolution algorithms, but there are only a few basic design decisions to be made by such systems when it comes to how to compare two individual instances. An analysis framework is presented that assesses the impact of different choices for these design decisions on coreference resolution in terms of precision, recall, and F-measure.