Text Similarity in Vector Space Models: A Comparative Study
This addresses the real-world problem of patent similarity modeling for researchers and practitioners, but it is incremental as it compares existing methods without introducing new ones.
The study tackled the problem of measuring semantic text similarity in natural language processing by evaluating various vector space models for patent-to-patent similarity, finding that TFIDF performs well for longer, technical texts or fine-grained distinctions, while more complex methods like neural models are only justified in specific cases.
Automatic measurement of semantic text similarity is an important task in natural language processing. In this paper, we evaluate the performance of different vector space models to perform this task. We address the real-world problem of modeling patent-to-patent similarity and compare TFIDF (and related extensions), topic models (e.g., latent semantic indexing), and neural models (e.g., paragraph vectors). Contrary to expectations, the added computational cost of text embedding methods is justified only when: 1) the target text is condensed; and 2) the similarity comparison is trivial. Otherwise, TFIDF performs surprisingly well in other cases: in particular for longer and more technical texts or for making finer-grained distinctions between nearest neighbors. Unexpectedly, extensions to the TFIDF method, such as adding noun phrases or calculating term weights incrementally, were not helpful in our context.