Comparison of a large number of texts (php + mysql)

Question

PHP

Good afternoon!
There was a task on one project to compare texts from a DB.
I tried similar_text but it turned out to be too simple for my task. I'll try to specify it.
I have, for example, 30 texts. They are divided into 5 categories. Task: compare all 30 to combine them into a smaller number of similarities. Language - php, base - mysql
What is the best way to use?

Answer 1

Try to compare words, not characters. In general, this is Bayesian categorization. But if you really have ~30 texts and not 30 thousand, then it's easier to do it by hand :)

Answer 2

you can look in the direction of Simple NaiveBayesClassifier for PHP
A good series of articles about Bayesian categorization in php:
Implement Bayesian inference using PHP
+
on Habré recently - Probabilistic models: Bayesian networks

Answer 3

If the number of classes at the output of the work is unknown, then, probably, Bayesian classification will not help you, and you need to look in the direction of clustering.