Litcius/Paper detail

Weather Prediction Using LSTM Neural Networks

Anitej Srivastava, S. Anto

20222022 IEEE 7th International conference for Convergence in Technology (I2CT)19 citationsDOI

Abstract

Weather comprises of components that are highly dynamic in nature and are also susceptible to extreme conditions with a changing frequency. The exact cause and effect relations among variables involved in weather forecasting are not tangible for the long range and they require some discovery. This makes prediction of weather in the distant future i.e., greater than 7-10 days a huge challenge. In this paper, we propose the use of Long Short Term Memory (LSTM) neural networks in order to predict the weather condition so that the model can better decide what to retain what to forget. Combining this with optimizations such as Gaussian and Median filtering has resulted in better accuracy in the long range as the model formed a much more informed pattern. LSTM has the capability to store key data that is fed remember it for the long term using gates.

Topics & Concepts

Computer scienceArtificial neural networkRange (aeronautics)Weather predictionWeather forecastingKey (lock)Artificial intelligenceLong short term memoryMachine learningTerm (time)Recurrent neural networkData miningMeteorologyComputer securityEngineeringPhysicsQuantum mechanicsAerospace engineeringMeteorological Phenomena and SimulationsHydrological Forecasting Using AIEnergy Load and Power Forecasting
Weather Prediction Using LSTM Neural Networks | Litcius