Implementation of the K- Nearest Neighbor and Neural Network for Predicting School Readiness to Enter Elementary School

Entering elementary school is a new experience for children, both those who have attended school at a previous level or who have never been. Some children have problems at school due to maturity and readiness. To do a psychological test too, it costs money and only a small proportion of parents can do it. The alternative was used as a web-based prediction system without psychology tests. This study aimed to predict the readiness of children entering elementary school. There were three decision classifications: ready; doubt, and not ready. The algorithms used were K-Nearest Neighbor and Neural Network to classify dataset and probe data. The variable was sourced from the profile of the child and his parents. Testing data were tested to obtain the highest accuracy. The results showed that neural networks had significantly better accuracy than K-Nearest Neighbor. It could be used as a recommendation in the prediction system. At least, the system was expected to be able to provide the right decisions, both for children, parents and the environment.