Tell me if a solution like this could work,
Let’s start right away with a procedure kindly offered by the dear 
ABG @Abraham_Getzler from the AppInventor 
community.
This is an effective method to compare two words, and in the topic there is a guide that explains step by step how it works.
Levenshtein distance
This is a compact version of what you did, the only extra thing I added is checking whether correct has been filled with text or whether the list similarWordLE5 is greater than similarWordLE3, if not, I show a message that the word is not there, otherwise I show some notifiers with the result.
Edit:
Since I didn’t really understand what data to enter, I just put random things
pear,apple,explanation1\n kiwi,ananas, \n coconut,banana,explanation3\n peanut,nut,explanation4
To adjust how precisely the words should be taken into consideration, you need to lower or raise these values: 5 and 3.
AIA
NoorProject.aia (9.0 KB)