How to verify if a word has valid syllables or not?

algorithm

Solution

Whenever we have collected a full pattern-string, we can either discard it and begin collecting to a new pattern-string, or keep it and try to get a longer pattern-string. We don't know in advance (without examining the rest of the input-string), whether we should keep or discard the current string, so we need to keep both possibilities in mind.

We can build a state machine that can keep track of this for us. The base-states are identified by the sequence of characters we have examined so far:

State   C           V
""      {"C"}       {"V",""}
"C"     {"CC"}      {"CV",""}
"CC"    {"CCC"}     {""}
"CCC"   {}          {""}
"CV"    {"CVC",""}  {}
"CVC"   {""}        {}
"V"     {""}        {}

Since we don't always know which action to take, we can be in several possible states at once. Those sets of possible states form super-states:

Index   Super-state        C    V
    0   {}                 0    0   Fail 
    1   {""}               2    9   Accept 
    2   {"C"}              3    8
    3   {"CC"}             4    1
    4   {"CCC"}            0    1
    5   {"","C"}           6   13   Accept
    6   {"C","CC"}         7    8
    7   {"CC","CCC"}       4    1
    8   {"","CV"}         12    9   Accept
    9   {"","V"}           5    9   Accept
   10   {"","C","CC"}     11   13   Accept
   11   {"C","CC","CCC"}   7    8
   12   {"","C","CVC"}    10   13   Accept
   13   {"","CV","V"}     12    9   Accept

The transitions are between super-states. Each member of the super-state is advanced with the same symbol. All members without such transition are discarded. If a member has two possible destinations, both are added to the new super-state.

You might notice that some rows are very similar. Super-state 3 and 7 have the same transitions. As are 6 and 11, and 8 and 13. You could collapse those into one state each, and update the indices. I'm not going to demonstrate that here.

This could easily be encoded into a programming language:

//                    index = 0  1  2  3  4   5  6  7   8  9  10  11  12  13
int[] consonant = new int[] { 0, 2, 3, 4, 0,  6, 7, 4, 12, 5, 11,  7, 10, 12 };
int[] vocal = new int[] {     0, 9, 8, 1, 1, 13, 8, 1,  9, 9, 13,  8, 13,  9 };
int[] accept = new int[] {    0, 1, 0, 0, 0,  1, 0, 0,  1, 1,  1,  0,  1,  1 };
int startState = 1;
int failState = 0;

bool CheckWord(string word)
{
    int state = startState;
    foreach (char c in word)
    {
        if (IsVocal(c))
        {
            state = vocal[state];
        }
        else if (IsConsonant(c))
        {
            state = consonant[state];
        }
        if (state == failState) return false;
    }
    return accept[state] != 0;
}

Example:

> CheckWord("pronunciation")
true

> CheckWord("pronunciationn")
false

Problem

I need to be able to process one word or words and verify that it has valid syllables. There are some syllabification rules that could be used: ``` V CV VC CVC CCV CCCV CVCC ``` where V is a vowel and C is a consonant. e.g., ``` pronunciation (5 Pro-nun-ci-a-tion; CCV-CVC-CV-V-CVC) ``` Or is there a simple code that can be used, or a library in c++? In class we're talking about binary search trees, hash tables, etc, but i can't really see the relation. Any help would appreciated, thanks.

Original source