Guerrini, Giordano (A.A. 2022/2023) Natural gas price prediction: a recurrent neural network based short-term forecast. Tesi di Laurea in Empirical finance, Luiss Guido Carli, relatore Giacomo Morelli, pp. 102. [Master's Degree Thesis]

PDF (Full text) Restricted to Registered users only Download (8MB) | Request a copy

Abstract/Index

Needings of natural gas forecasting models. Solution proposed by the paper. Methodology. Time series dispersion. Stationarity and variance. Sample entropy and dispersion entropy. Neural networks. Feedforward neural networks. Simple RNN. GRU. LSTM. Network type selection. Keras library. Keras API. Keras layers: sequential and functional. Keras preprocessing tools. Keras optimization algorithms. Loss function selection. Keras tool for computational costs curtailing. Benchmark models. ANN benchmark. SRNN benchmark. LSTM benchmark. Univariate and multivariate procedures. Univariate input tensor. Multivariate input tensor. Refining process. Hyperparameters tuning program. Python code for hyperparameters optimisation. Weights optimization process. Models comparison. MSE. RMSE. MAE. MAPE.

References

Bibliografia: pp. 77-81.

Thesis Type:	Master's Degree Thesis
Institution:	Luiss Guido Carli
Degree Program:	Master's Degree Programs > Master's Degree Program in Economics and Finance (LM-56)
Chair:	Empirical finance
Thesis Supervisor:	Morelli, Giacomo
Thesis Co-Supervisor:	Peracchi, Franco
Academic Year:	2022/2023
Session:	Autumn
Deposited by:	Alessandro Perfetti
Date Deposited:	24 May 2024 14:27
Last Modified:	24 May 2024 14:27
URI:	https://tesi.luiss.it/id/eprint/38680

Downloads

Repository Staff Only

View Item