# Text Similarity in Vector Space Models: A Comparative Study

arXiv: https://arxiv.org/abs/1810.00664 · HTML: https://mlsanity.com/paper/1810.00664/

## Key facts

- Cited by 13 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 24 Sep 2018.

## Built on by (most cited first, 13 of 13)

- [Data-Juicer: A One-Stop Data Processing System for Large Language Models](https://mlsanity.com/paper/2309.02033/) (2023)
- [The Counting Power of Transformers](https://mlsanity.com/paper/2505.11199/) (2025)
- [Evaluating Metrics for Bias in Word Embeddings](https://mlsanity.com/paper/2111.07864/) (2021)
- [SciEv: Finding Scientific Evidence Papers for Scientific News](https://mlsanity.com/paper/2205.00126/) (2022)
- [Text Classification using Graph Convolutional Networks: A Comprehensive Survey](https://mlsanity.com/paper/2410.09399/) (2024)
- [The SAME score: Improved cosine based bias score for word embeddings](https://mlsanity.com/paper/2203.14603/) (2022)
- [Location retrieval using visible landmarks based qualitative place signatures](https://mlsanity.com/paper/2208.00783/) (2022)
- [Tutorial Recommendation for Livestream Videos using Discourse-Level Consistency and Ontology-Based Filtering](https://mlsanity.com/paper/2209.04953/) (2022)
- [Machine Learning and Statistical Approaches to Measuring Similarity of Political Parties](https://mlsanity.com/paper/2306.03079/) (2023)
- [Modeling Political Orientation of Social Media Posts: An Extended Analysis](https://mlsanity.com/paper/2311.12323/) (2023)
- [An Iterative Approach to Topic Modelling](https://mlsanity.com/paper/2407.17892/) (2024)
- [Comparative Analysis of Document-Level Embedding Methods for Similarity Scoring on Shakespeare Sonnets and Taylor Swift Lyrics](https://mlsanity.com/paper/2412.17552/) (2024)
- [From Divergence to Consensus: Evaluating the Role of Large Language Models in Facilitating Agreement through Adaptive Strategies](https://mlsanity.com/paper/2503.15521/) (2025)

## Abstract (arXiv)

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.