Development of Animated Science Learning Videos Based on Deep Learning to Improve Student Learning Outcomes

Abstrak

This study aims to improve student learning outcomes through animation-based learning videos with a deep learning approach. Conventional teaching methods are less effective for science lessons, resulting in less engagement and achievement of student learning outcomes. The integration of deep learning implemented through animated videos has an impact on teacher success in learning through visualization, adaptive, automatic narration, and intelligent synchronization, making lessons interactive. This study uses the ADDIE model development method. Data collection is based on expert validation, teacher and student responses, and pretest posttest tests. The results show that the level of student learning success through animated videos meets very good feasibility standards, with an average score of 92% for media experts and 89% for material experts. Student responses increased by an average of 91%. Animated videos in the implementation process are very effective with a score from the pretest results of 68.3 to the posttest increasing to 86.7 with an N-Gain score of 0.58 in the moderate category. Based on learning outcomes during the implementation process, animated videos for science lessons are feasible and interesting, but also effective in improving student learning outcomes, offering innovative digital resources to support meaningful science education