sample java code for approximate string matching or boyer-moore extended for approximate string matching

Question

I need to find 1.mismatch(incorrectly played notes), 2.insertion(additional played), & 3.deletion (missed notes), in a music piece (e.g. note pitches [string values] stored in a table) against a reference music piece.

This is either possible through exact string matching algorithms or dynamic programming/ approximate string matching algos. However I realised that approximate string matching is more appropriate for my problem due to identifying mismatch, insertion, deletion of notes. Or an extended version of Boyer-moore to support approx. string matching.

Is there any link for sample java code I can try out approximate string matching? I find complex explanations and equations - but I hope I could do well with some sample code and simple explanations. Or can I find any sample java code on boyer-moore extended for approx. string matching? I understand the boyer-moore concept, but having troubles with adjusting it to support approx. string matching (i.e. to support mismatch, insertion, deletion).

Also what is the most efficient approx. string matching algorithm (like boyer-moore in exact string matching algo)?

Greatly appreciate any insight/ suggestions. Many thanks in advance

Answer 1

here, see the first (Boyer.java) of this results page

Answer 2

You could start with the Wikipedia page on approximate string matching.

The problem is that this is a complex field, and simply looking at / copying some example code probably won't help you understand what is going on.

EDIT - besides, I don't see how Boyer-Moore would adapt to approximate string matching.

Answer 3

Here is the C# Boyer-More code, which can be tweeked to BMH or approximate matching.

Dictionary<char, int> ShiftSizeTable = new Dictionary<char, int>();

        //Calculate Shifit/Skip count for each element in pattern text. So that we can skip that many no of Characters in given text while searching.
        public void PreProcessBMSBadMatchTable(char[] patternCharacters)
        {
            ShiftSizeTable.Clear();

            int totalCharacters = patternCharacters.Length;

            for (int lpIndex = 0; lpIndex < totalCharacters; lpIndex++)
            {
                //Calculate the shift size for each character in the string or char array.
                int ShiftSize = Math.Max(1, (totalCharacters - 1) - lpIndex);

                //If the charater is already exists in the ShiftSize table then replace it else add it to ShiftSize table.
                if (ShiftSizeTable.ContainsKey(patternCharacters[lpIndex]))
                {
                    ShiftSizeTable.Remove(patternCharacters[lpIndex]);
                }

                ShiftSizeTable.Add(patternCharacters[lpIndex], ShiftSize);
            }
        }

        //Use the PreProcessed Shift/Skip table to find the pattern Characters in text and skip the bad Characters in the text.
        public int BoyerMooreSearch1UsingDictionary(char[] textCharacters, char[] patternCharacters)
        {        
            PreProcessBMSBadMatchTable(patternCharacters);

            int SkipLength;
            int patternCharactersLenght = patternCharacters.Length;
            int textCharactersLenght = textCharacters.Length;

            // Step2. Use Loop through each character in source text use ShiftArrayTable to skip the elements.
            for (int lpTextIndex = 0; lpTextIndex <= (textCharactersLenght - patternCharactersLenght); lpTextIndex += SkipLength)
            {
                SkipLength = 0;

                for (int lpPatIndex = patternCharactersLenght - 1; lpPatIndex >= 0; lpPatIndex--)
                {
                    if (patternCharacters[lpPatIndex] != textCharacters[lpTextIndex + lpPatIndex])
                    {
                        SkipLength = Math.Max(1, lpPatIndex - ShiftSizeTable[patternCharacters[lpPatIndex]]);
                        break;
                    }
                }
                if (SkipLength == 0)
                {
                    return lpTextIndex;    // Found
                }
            }
            return -1; // Not found
        }