Topical Segmentation: a Study of Human Performance and a New Measure of Quality

In a large-scale study of how people find topical shifts in written text, 27 annotators were asked to mark topically continuous segments in 20 chapters of a novel. We analyze the resulting corpus for inter-annotator agreement and examine disagreement patterns. The results suggest that, while the overall agreement is relatively low, the annotators show high agreement on a subset of topical breaks – places where most prominent topic shifts occur. We recommend taking into account the prominence of topical shifts when evaluating topical segmentation, effectively penalizing more severely the errors on more important breaks. We propose to account for this in a simple modification of the windowDiff metric. We discuss the experimental results of evaluating several topical segmenters with and without considering the importance of the individual breaks, and emphasize the more insightful nature of the latter analysis.