Prototype Permainan Cerdas HTML 5 Berbasis Artificial Neural Network (ANN)

Abstract

Multimedia is currently widely used as a medium of learning, especially in kindergarten. As in educational games that are often found today, they have a passive nature, which only clicks every button on the game so there is no difference. In this study, the use of an artificial network is used as a knowledge center for artificial intelligence in the guessing fruit game. The right artificial network to become the center of artificial intelligence because the artificial intelligence artificial network that continues to evolve is getting smarter. To overcome this problem, an artificial neural network (ANN) algorithm is needed to make educational games smarter. The game uses HTML as the user interface. The test method uses the Confusion Matrix test which can provide precision information, recall, fi-score, and Black Box Testing to see that the functions of the buttons in the game can work properly.  Describing the accuracy of how accurate the model is in classifying correctly. In this study, an accuracy value of 97% means that the model can classify fruit images well. During the image prediction process, it takes quite a long time, this is because the model recognizes the pattern/shape of each type of fruit. The more types of fruit used, the longer the model takes to predict.