New paper: clearer topic profiles for large collections of laws
I am happy to share our new paper in Applied Sciences, Ranked Multi-Label-Augmented Topic Modeling for Legislative Content Profiling.
Large collections of laws are difficult to explore. Topic modeling can reveal their broad themes, but a topic alone is often too general to describe the content of an individual law.
In this work, we combine topic modeling with ranked, descriptive labels. The result is a clearer profile for each law: the topic gives the overall context, while the labels explain the most relevant details. We applied the approach to Italian legislation to make large legislative collections easier to browse and understand.