The estimation of cartoons effect on children´s behavior based on exaggeration action by using neural network

Exaggeration is one of 12 basic principles of animation. The aim of exaggeration is to increase audience´s significance or attention. Unfortunately, exaggeration action can be in the form of violence scenes. So, it may affect the child´s psychology and behavior. This research aimed to make prediction system of cartoons which could have negative impact on children by using neural network method. The input parameters of our neural network were scenes duration, duration of exaggeration action, total duration of actions, and quantitative of exaggeration action. The output parameter of our neural network system are G, PG, PSG and R. These labels are MPPA rating value. Our neural network prediction system used three scenarios of parameters input set with back propagation method, the value of epoch: 0, number of epoch: 500, learning rate: 0.3 and momentum: 0.2. These scenarios were qualitative, quantitative and the combination of quantitative-quantitative parameters input set. The accuracy of our prediction system using quantitative parameters input set was 76%, second scenario got 63% accuracy rate and third scenario got 76% accuracy rate. This result showed that the exaggeration scene in an animation film is able to influence the behavior of children despite the scene have short duration.