Text Similarity in Vector Space Models: A Comparative Study

TL;DR

This study compares various vector space models for semantic text similarity, finding that traditional TFIDF often outperforms neural embeddings in complex, technical, or detailed text scenarios, with added computational costs not always justified.

Contribution

The paper provides a comprehensive comparison of TFIDF, topic models, and neural models for patent-to-patent similarity, highlighting the conditions where each method excels or falls short.

Findings

TFIDF performs well on longer, technical texts

Neural models are justified only for condensed texts with trivial similarity tasks

Extensions to TFIDF did not improve performance in this context

Abstract

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.