SMART WEATHER FORECAST USING ARTIFICIAL NEURAL NETWORK

Abstract

Smart Weather forecast using deep learning based Nodemcu is the system that can analyze the weather in a next minute and notify the user based on the result of training system using ANN (Artificial Neural Network). The training system in ANN is using error based case as the improvement of the forecast outcome to be more accurate. The system parameters determine the training’s material in ANN to predict the weather. The first parameter is water which output the result of prediction and the next to last parameters are humidity and temperature. The data will be processed and calculated in the server and return the result prediction to the user email or phone. There are particular events that is relying the weather to be considered as having mutual relation. Such as music concert, outdoor sport and any other outdoor activities. Especially many peoples in Indonesia who is riding a bike and relying on a good weather. The system also helps the housewife to monitor their clothesline outside the house. In conclusion, the specific function of this weather forecast system branches the mutual relation for any fields that relying a good weather. Stages of research starts from data collection, data training, analysis, hardware design, implementation, testing and finalization of the system. System design is using UML modeling and display implementation using state machine diagram, as well as system testing using Blackbox testing method and data accuracy. The end result of this study specifically the system can be used as a notification to alert people that the rainy will come.