In most cases, making the regex rules smarter by adding new code logic or libraries will be required. One such way to do this is by using the concepts behind normalizing your text called stemming and lemmatization. Both terms are rooted in the study of linguistics, and how they are adopted to be used in technology has exploded due to integrating NLP solutions into everything, from customer service to speech-to-text features.
When applied to NLP, stemming is when any word is programmatically identified to its common root form. In this process, any suffix, plural form, or synonym that exists for the word is identified. Stemmers require a reference dictionary or lookup to be accurate, so the source language is required. Lemmatization takes into account all of the variations of a word so it can be rooted back to a dictionary source. From my research, both stemming and lemmatization are used together in NLP and you can start by using the open source...
