Phonetic Analysis

In this chapter we've looked at traditional encoding methods and also hash generation, but Codec offers an additional set of classes for use in phonetic analysis.

It's beyond the scope of this text to delve into the theory behind Soundex and Metaphone phonetic analysis, but suffice to say that these algorithms provide different mechanisms for generating a phonetic key. This key is designed to be able to help answer the question, "How much does this word sound like another?"

Table 12-1 shows how different phonetic algorithms included with Codec translate words into phonetic keys. Depending on the application you are trying to build, different algorithms may be of interest. For example, a spell-checking application might use Metaphone to determine alternative words as part of a spell-checking routine, whereas a voice-recognition system might rely on the RefinedSoundEx algorithm to determine the most likely words the speaker is trying to convey. Notice that while SoundEx produces the same length key for all words, the other algorithms produce variable length keys.

Table 12-1. Phonetic Algorithm Results

Word	SoundEx	RefinedSoundEx	Metaphone	Double Metaphone
hello	H400	H070	HL	HL
fellow	F400	F2070	FL	FL
mellow	M400	M8070	ML	ML
monster	M523	M8083609	MNST	MNST
monstrous	M523	M80836903	MNST	MNST

By using these algorithms and a very large set of words, you can generate a set of keys linking the words phoneticallythe fundamentals needed for a spell checker.

Listing 12-10 shows the code used to generate the terms as shown in Table 12-1.

Listing 12-10. Phonetic Key Generation Source

 public static void phoneticDemo() {     printHeader("Phonetic Demo");     String[] words =         { "hello", "fellow", "mellow", "monster", "monstrous" };     System.out.println("Notice the sounds of the words, and");     System.out.println("how the sounds are translated into 4");     System.out.println("character character flags.");     System.out.print("Word, SoundEx, RefinedSoundEx, ");     System.out.println("Metaphone DoubleMetaphone");     for (int i = 0; i < words.length; i++)     {         System.out.print(words[i] + ", ");         System.out.print(new Soundex().encode(words[i]) + ", ");         System.out.print(             new RefinedSoundex().encode(words[i]) + ", ");         System.out.print(             new Metaphone().encode(words[i]) + ", ");         System.out.print(             new DoubleMetaphone().encode(words[i]));         System.out.println();     } } 

For more information on phonetic analysis, see the following sites:

http://encyclopedia.thefreedictionary.com/Soundex

http://www.archives.gov/research_room/genealogy/census/soundex.html

http://aspell.sourceforge.net/metaphone/