php - find segments of similar text within a string with php - php

Question

I am familiar with the similar_text() function in PHP but what I am trying to think of how to do is find potential links I could add in my content to other articles that I have.
What I want is to be able to scan through all my content in each post and find a segment of text in a post that is similar to a title of another post.
So lets say I have the following structure
$contentfromapost = "this is example text of the content in this article. It talks about things that people like to do for fun and vacation spots where they can do them all around the world"
$titleofpost1 = "Yellow cats are fun to throw in the snow"
$titleofpost2 = "Vacation Rentals in Fun parts of the world"
So my idea is to scan the first post content, then scan the titles of all my articles.
As you can see in my example $titleofpost2 has matching keywords to the $contentfromapost.
Then I would want to be able to grab the segment of text in $contentfromapost and send a link to that post with the similar title. I would maybe use anchor text "fun and vacation spots where they can do them all around the world" to link to that second post.
I want to build this to help me find other posts that I could link too within an article. Potentially I would like it to be able to automatically add the link with the section of text.
Anyways trying to see how I could structure this, any ideas would help

Answer 1

Here is my suggestion on this. This might be helpful.
You can make in two steps :
First Step (Data collection)
1) You can make a table which will contain the keywords and PostId (Here I am assuming each post will have a unique ID) along with other required fields.
2) For making keyword list you can parse each post and filter out Nouns, Verbs.
3) For each keyword and Postid will have a mapping in table.
Second Step (Data Extraction)
Now suppose you have to find the similar post of POST-1
1) First find the all keyword stored for POST-1. And store it in a data structure
2) Now take each keyword and find it in Table which do not belongs to POST-1. Now you will have a Data Structure of PostId for each keyword. Since one postid can come multiple times for a Keyword so here we can also introduce weight of each postid.
3) Now suppose you have 5 keyword for Post-1. Means you have 5 Data structure of Postid with their weight.
4) Now combine all Data Structures. During combination you will have to take care that if any PostId is coming more time means we are also increasing the weight.
5) Finally you will get the PostId which will have most weight. This will be close to your Post-1.
Hope it will make sense.