First level text prediction using data mining and letter matching in IEEE 802.11 mobile devices

When a user uses an 802.11 wireless mobile device like a Personal Digital Assistant (PDA) to type in a message, typing every word fully can be quite tedious. This paper investigates into a first level text prediction scheme through data mining and letter matching which could help the user with writing messages using predictive approach, to be used for short message service text or for email text. The performance results of simulation show that the results are quite encouraging and the accuracy of first level text prediction is above 80% for a huge sample, where a given word is mapped to a subset of possible next words, in fixed and flexible scenarios. Though the prediction is tried with English language, it can be applied to other similar languages.