Adaptable probabilistic flick keyboard based on HMMs

To provide an accurate and user-adaptable software keyboard for touchscreens, we propose a probabilistic flick keyboard based on HMMs. This keyboard can reduce the input error by taking the time series of the actual touch position into consideration and by user adaptation. We evaluated performance of the HMM-based flick keyboard and MLLR adaptation. Experimental results showed that a user-dependent model reduced the error rate by 28.2%. In a practical setting, MLLR user adaptation with only 10 words reduced the error rate by 16.5% and increased typing speed by 10.5%.