Fine-Tunning a MAP Error Corrections Algorithm for Five-Key Chording Keyboards

Different typing devices lead to different typing error patterns. In addition, different persons using the same device have different error patterns. Considering this, we propose and evaluate a spelling algorithm specifically designed for a five-key chording keyboard. It uses the maximum a posteriori probability rule, the probabilities that one character is typed for another, named confusion probabilities, and a dictionary model. Our study shows that the proposed algorithm reduces the substitution error rate from 7.60% to 1.25%. In comparison, MsWord and iSpell reduce the substitution error rates to 3.12% and 3.94%, respectively. The error rate can be further reduced to 1.15% by using individual confusion matrices for each user.