Prediction of Smarthphone Charging using K-Nearest Neighbor Machine Learning
Faza Ghassani, Maman Abdurohman, Aji Gautama Putrada · 2018 Third International Conference on Informatics and Computing (ICIC) · 2018
This paper proposes smartphone charging system using kNN for increasing charging time accuracy. Smartphone charging is done every time to ensure the battery is fully charged. The smart phone user's habits lead to decreased in battery capacity and battery life faster than it should. Stopping the charging cycle on time is required to avoid decreasing capacity and battery life due to overcharging. Charging predictions are performed or stopped by viewing the state of charge and timestamp periodically that are sent over from the smartphone and processed using the k-Nearest Neighbor algorithm. The smartphone will stop charging when the prediction of k-Nearest Neighbor gets the state of charge and the timestamp in seconds according to the user's habit of getting the 100 percent state of charge and the timestamp. Based on experiments have been done, the result show, K-nearest neighborhood machine learning algorithm can predict the charging decision to be continued or stopped and in this case, K = 2 is the optimal K because the F1-Score is close to 1 and higher F1-Score (0.78) compared with other K.