Morphological Analysis of Speech Translation into Indonesian Sign Language System (SIBI) on Android Platform

The main problem that occurs with people with hearing impairment is the difficulty of communicating, both among fellow deaf and non-hearing impaired. This difficulty is caused by not all non-hearing impairments have skills in using sign language such as the Sistem Isyarat Bahasa Indonesia (SIBI). The presence of a SIBI dictionary cannot be used practically and less effective if people want to translate sentences to sign language. The goal that will be achieved in this research is to create assistive applications that can change the speech in form of sentences into SIBI so that it can facilitate communication between people with hearing impairment and people with non-hearing impairment. The specific target to be achieved in this research is the creation of an android application that can change the speech in form of sentences into SIBI using morphological analysis with modification of the Enhanced Confix Stripping (ECS) stemming algorithm. The method used consists of literature studies and requirement analysis, Android Speech API integration, parsing the sentence (through morphological analysis with modification of the ECS stemming algorithm), creating SIBI video datasets, testing algorithms, and testing overall. The results showed that the average accuracy of the Android Speech API is 94.06%, the average accuracy rate of morphological analysis with modification of the ECS stemming algorithm is 95%, and the overall level of conformity is 80,71%. These results indicate that overall the speech translator into SIBI is working very well.