Implementation Analysis of KEBI Application on Non-Standard Words Detection in Student Essay Tasks on Language Variety

Non-standard word errors in writing sometimes escape the attention of writers or editors due to several factors, including unintentional or limited time in editing. Non-standard word is one of the Indonesian spelling error. Therefore, the application for detecting non-standard words in writing, KEBI, was created to minimize the height of non-standard words in writing. In this study, KEBI was analyzed in terms of application implementation to five various language essay documents as student assignments for Indonesian language courses. There are three things that are analyzed, namely the interface page, processing time, and the level of accuracy of detection and correction using the precision and recall method (confusion matrix). The overall research method includes literature study, data collection, testing, measurement of display, time, and accuracy. Based on the results of the analysis, the interface page display shows good quality. The processing time required by KEBI to detect non-standard words is in the range of 0.02-0.06 seconds. Meanwhile, KEBI's non-standard word detection accuracy rate is 86%, while KEBI's non-standard word correction accuracy rate is 90%.