├── img ├── fourier.gif ├── ND_model.png ├── ND_structure.png ├── ND_comparisson.png ├── ND_structure_formula.png └── types_of_forecasting.png └── README.md /img/fourier.gif: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/trokas/neural_decomposition/HEAD/img/fourier.gif -------------------------------------------------------------------------------- /img/ND_model.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/trokas/neural_decomposition/HEAD/img/ND_model.png -------------------------------------------------------------------------------- /img/ND_structure.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/trokas/neural_decomposition/HEAD/img/ND_structure.png -------------------------------------------------------------------------------- /img/ND_comparisson.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/trokas/neural_decomposition/HEAD/img/ND_comparisson.png -------------------------------------------------------------------------------- /img/ND_structure_formula.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/trokas/neural_decomposition/HEAD/img/ND_structure_formula.png -------------------------------------------------------------------------------- /img/types_of_forecasting.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/trokas/neural_decomposition/HEAD/img/types_of_forecasting.png -------------------------------------------------------------------------------- /README.md: -------------------------------------------------------------------------------- 1 | # Neural Decomposition of Time-Series Data 2 | 3 | We will look at recent Gashler neural network technique for the analysis and extrapolation of time-series data called Neural Decomposition (ND). ND outperforms popular time-series forecasting techniques including LSTM, echo state networks, ARIMA and SARIMA. 4 | 5 | *Artificial Intelligence Group Meetup \#4, 2018-01-24* 6 | 7 | --------------------------------------------------------------------------------