String Approximation (Fetching the nearest matched String from Dictionary)

I think it's better to use lucene library it has a package called org.apache.lucene.search.spell you can use it easily. it provide 3 algorithms NGramDistance, LevensteinDistance, JaroWinklerDistance. try this

You can calculate a Levenshtein distance between your String and strings in your dictionary to find the closest matches. This may not be the best for spell checking as it gives no favour to letter being swapped around or words which are phonetically similar. e.g. question is closer to resting than kwizchum.

For more examples, you can read up on http://en.wikipedia.org/wiki/Approximate_string_matching

I just wanted to add that StringUtils also has a convenient Levenshtein Distance method since version 3.0

public static int getLevenshteinDistance(CharSequence s,
                     CharSequence t)

After that it's just as simple as iterating through the collection and remembering the closest match:

public static Object findClosestMatch(Collection<?> collection, Object target) {
    int distance = Integer.MAX_VALUE;
    Object closest = null;
    for (Object compareObject : collection) {
        int currentDistance = StringUtils.getLevenshteinDistance(compareObject.toString(), target.toString());
        if(currentDistance < distance) {
            distance = currentDistance;
            closest = compareObject;
        }
    }
    return closest;
}

Note that the method above does require collection to be null-safe and toString() to be sensfully implemented.

You can try Levenshtein Distance techinque.

The simple idea you have four basic operations:

• Insertion (hell -> hell o)
• Replacement (nice -> r ice)
• Deletion (bowlin g -> bowlin)
• Swapping (brohter -> bro th er)

You algorithm should calculate the distance between your word and every word in dictionary. The smallest distance means that this word matches more accurate with given input.