An AI based approach to evaluate students old script writings

Abstract

Abstract Preserving cultural heritage, such as old manuscripts as expressions of human communities is one of sustainable development goals (SDGs) targets. In Indonesia, old manuscripts written in Batak language include many valuable insights for new generation. The manuscripts recorded variety of contents from folklore to medicinal prescriptions. Nevertheless, those heritages are worried to be vanishing in face of information technology advancement. Due to this reason, some efforts have been made by the Indonesian, for instance teaching the writing of the old language at schools in rural areas where this language is used. In this research, an experiment was conducted to recognize the writings of Batak scripts among the junior high school students who belong to Batak tribe in Balige, which is one of the rural districts of Indonesia. A group of students were participating to write some short phrases using digital graphic drawing pen tablets via a web-based application. Recognition of written scripts for the group was performed using convolutional neural network (CNN). The objective of this research is to evaluate capability of students' writings of Batak scripts using digital technology. The aims are two folds: to preserve the old scripts for new generations by enhancing their attention to the cultural heritage and to promote the quality of education using artificial intelligence techniques. The experimental results show that using digital tools students are facing some difficulties in coping with digital technology, whereby they were not able to finish the tasks in time frame given and showed lower accuracy in terms of phrases recognition. Nevertheless, students show enthusiasm on the usage of digital technology in the writing process and consider that preserving the cultural heritage is a necessity.