Performance Evaluation of MissLess Soft Keyboard with Recommendation

In this paper, we evaluated the performance of recommendation and run time of the MissLess soft keyboard developed in the previous research. The MissLess keyboard assigns a hash code per each word for all words within its mobile dictionary. It decides recommendation words through three consecutive processes such as hash filtering, sorting based on spelling similarity, and finally recommendation based on frequency of use. Each process has some factors to have an impact on the recommendation success. We conducted experiments in an Android mobile device running the MissLess keyboard and measured performance of recommendation and run time overhead according to the impact factors. In this paper, we showed the experiment results.