Application of Artificial Neural Network (ANN) to improve forecasting of sea level

Comparison of artificial neural networks and harmonic analysis for sea level forecasting (Urias coastal lagoon, Mazatlan, Mexico)

“…Primero, se realizó un análisis armónico (HA, por sus siglas en inglés) (Foreman y Henry 1989, Salas-Pérez et al 2008: mediante un análisis espectral de Fourier, se analiza la serie de tiempo y las frecuencias obtenidas (componentes de mareas) se usan para estimar los valores futuros. El HA es una técnica poderosa y ampliamente utilizada para predecir las variaciones del nivel del mar, pero tiene las siguientes desventajas (Lee y Jeng 2002, Filippo et al 2012: (i) no considera variaciones locales inducidas por el forzamiento meteorológico, lo cual puede conducir a importantes errores de predicción; y (ii) requiere series de datos relativamente largas (~1 año) para poder estimar suficientes armónicos.…”

“…In this approach, one defines a dynamical model that requires a set of initial conditions, which are past values of the time series, used to predict future values. Applications of ANN for sea level forecasting have been reported by several authors (Lee and Jeng 2002, Salas-Pérez et al 2008, Filippo et al 2012, Shetty and Dwarakish 2013. Although NAR networks are versatile modeling tools, they also present drawbacks, such as their sensitivity to data quality and the need of preliminary tests to set up an adequate network.…”

Prediction of refractive index of binary solutions consisting of ionic liquids and alcohols (methanol or ethanol or 1-propanol) using artificial neural network

“…The ANN algorithm has been used extensively for several applications, e.g. electronics [9,26,32] , forecasting [2,10,28] , pharmaceutical research [1] , financial institutions [39] , etc. but limited use of such algorithm has been made for chemical engineering applications.…”

Tidal-Level Forecasting Using Artificial Neural Networks Along the West Coast of India

“…Feed forward neural network with Resilient Back Propagation (RBP) learning algorithm was used to predict tide levels and supplement missing data with quicker computation 16) . To study the effect of meteorological parameters of wind speed and sea-level atmospheric pressure on tidal predictions 17) , incorporated threehour wind speed data and atmospheric pressure data in the input, along with calculated tide data from the harmonic analysis method for the network training. The importance of the meteorological parameters was highlighted by the results obtained, which showed a considerable decrease in prediction error when additional inputs were provided.…”