Segmenting words using scaled probabilities

A system comprising:

• a dictionary including n-grams and corresponding probabilities of each n-gram identifying a word;
• a scaling engine including scaling weights corresponding to each n-gram, the scaling weights depending on the number of atomic units in each n-gram;
• and
• scaled probabilities of each n-gram identifying a word, wherein the scaled probabilities of each n-gram identifying a word is determined including applying a scaling weight to a corresponding probability for each n-gram identifying a word.

A computer program product, tangibly stored on a computer-readable medium, comprising instructions operable to cause a programmable processor to:

• receive a probability of a n-gram identifying a word;
• determine a number of atomic units in the corresponding n-gram;
• identify a scaling weight depending on the number of atomic units in the n-gram;
• and
• apply the scaling weight to the probability of the n-gram identifying a word to determine a scaled probability of the n-gram identifying a word.

A system comprising:

• means for receiving a probability of a n-gram identifying a word;
• means for determining a number of atomic units in the corresponding n-gram;
• means for identifying a scaling weight depending on the number of atomic units in the n-gram;
• and
• means for applying the scaling weight to the probability of the n-gram identifying a word to determine a scaled probability of the n-gram identifying a word.