Many times, common terms like coronavirus, COVID-19, Omicron, and Delta virus could be treated the same while measuring the sentence similarity.
Instead of building a thesaurus, which is not a trivial task, a set of relevant words such as COVID-19, vaccine, and CDC are saved in a database.
The database is checked whenever a similarity measurement runs.
However, this approach is a temporary fix because it misses many words.
Researchers can consider taking advantage of online services like https://www.synonym.com/, which provides a list of synonyms of a word, but it slows down the execution tremendously.
On the other hand, saving all synonyms in a database is not feasible since it would take much space from a database.
WordNet (A Lexical Database for English)
Instead of building a thesaurus yourself, the WordNet could be used to facilitate synonym matching.
The Wordnet has become one of the most popular dictionaries for use in natural language processing (NLP) and other areas of language technologies.
This is primarily due to their structure as a graph of words, that is much easier for computers to understand than the traditional form of a dictionary.