You're asking for the "best" algorithm to learn an sequence of strings following an arbitrary rule. Your problem is that the non-trivial rules are external to the data. Unless you have multiple users applying the same sequence, you have no way of "learning" a sequence to apply later. For instance, given the obvious month sequence you listed, from where is your prediction model expected to learn the sequence of month names?
As stated, this isn't a Machine Learning problem ... yet. The emphasis is on the "learning". You need "enough" users employing a sequence of month names to learn that this is a recurring sequence. Without enough training data, you don't have the material needed to teach your program.
manpreet![Tuteehub forum best answer]()
Best Answer
2 years ago
I need to be able to predict what a users password is next likley to be, given a list of historic passwords they have used before.
For example let's say I have used the passwords: Password1, Password2, Password3. You could take a good guess that my next password is likely to be Password4. Or given January2017!, March2017!, May2017!, July2017! you could guess my next password would be September2017!.
The number of historic passwords I have could be as few as 2, but as many as ~50. The use case here is to prevent them from choosing a password that is "predicatable" and use it as a training aid.
Is Machine Learning the best solution here? I'm guessing not as my "training data" would be very limited per user. Guessing number increases like in my first example would be fairly straight forward and I could write a simple algorithm. But what about +2 month password? I don't want to write in all edge cases of passwords.
Any suggestions?