All Projects → Meatssauce → SCINet

Meatssauce / SCINet

Licence: other
Forecast time series and stock prices with SCINet

Programming Languages

python
139335 projects - #7 most used programming language
Jupyter Notebook
11667 projects

Projects that are alternatives of or similar to SCINet

Gluon Ts
Probabilistic time series modeling in Python
Stars: ✭ 2,373 (+8375%)
Mutual labels:  time-series, forecasting
Modeltime
Modeltime unlocks time series forecast models and machine learning in one framework
Stars: ✭ 189 (+575%)
Mutual labels:  time-series, forecasting
Pyfts
An open source library for Fuzzy Time Series in Python
Stars: ✭ 154 (+450%)
Mutual labels:  time-series, forecasting
notebooks
Code examples for pyFTS
Stars: ✭ 40 (+42.86%)
Mutual labels:  time-series, forecasting
Deep XF
Package towards building Explainable Forecasting and Nowcasting Models with State-of-the-art Deep Neural Networks and Dynamic Factor Model on Time Series data sets with single line of code. Also, provides utilify facility for time-series signal similarities matching, and removing noise from timeseries signals.
Stars: ✭ 83 (+196.43%)
Mutual labels:  time-series, forecasting
Statespacemodels.jl
StateSpaceModels.jl is a Julia package for time-series analysis using state-space models.
Stars: ✭ 116 (+314.29%)
Mutual labels:  time-series, forecasting
Introduction To Time Series Forecasting Python
Introduction to time series preprocessing and forecasting in Python using AR, MA, ARMA, ARIMA, SARIMA and Prophet model with forecast evaluation.
Stars: ✭ 173 (+517.86%)
Mutual labels:  time-series, forecasting
Msgarch
MSGARCH R Package
Stars: ✭ 51 (+82.14%)
Mutual labels:  time-series, forecasting
ForestCoverChange
Detecting and Predicting Forest Cover Change in Pakistani Areas Using Remote Sensing Imagery
Stars: ✭ 23 (-17.86%)
Mutual labels:  time-series, forecasting
query-selector
LONG-TERM SERIES FORECASTING WITH QUERYSELECTOR – EFFICIENT MODEL OF SPARSEATTENTION
Stars: ✭ 63 (+125%)
Mutual labels:  time-series, forecasting
Forecastml
An R package with Python support for multi-step-ahead forecasting with machine learning and deep learning algorithms
Stars: ✭ 101 (+260.71%)
Mutual labels:  time-series, forecasting
CoronaDash
COVID-19 spread shiny dashboard with a forecasting model, countries' trajectories graphs, and cluster analysis tools
Stars: ✭ 20 (-28.57%)
Mutual labels:  time-series, forecasting
Anticipy
A Python library for time series forecasting
Stars: ✭ 71 (+153.57%)
Mutual labels:  time-series, forecasting
Forecasting
Time Series Forecasting Best Practices & Examples
Stars: ✭ 2,123 (+7482.14%)
Mutual labels:  time-series, forecasting
Neural prophet
NeuralProphet - A simple forecasting model based on Neural Networks in PyTorch
Stars: ✭ 1,125 (+3917.86%)
Mutual labels:  time-series, forecasting
Java Timeseries
Time series analysis in Java
Stars: ✭ 155 (+453.57%)
Mutual labels:  time-series, forecasting
Informer2020
The GitHub repository for the paper "Informer" accepted by AAAI 2021.
Stars: ✭ 771 (+2653.57%)
Mutual labels:  time-series, forecasting
Pmdarima
A statistical library designed to fill the void in Python's time series analysis capabilities, including the equivalent of R's auto.arima function.
Stars: ✭ 838 (+2892.86%)
Mutual labels:  time-series, forecasting
dbnR
Gaussian dynamic Bayesian networks structure learning and inference based on the bnlearn package
Stars: ✭ 33 (+17.86%)
Mutual labels:  time-series, forecasting
AutoTS
Automated Time Series Forecasting
Stars: ✭ 665 (+2275%)
Mutual labels:  time-series, forecasting

SCINet: Sample Convolution and Interaction Networks Implementation (Work in Progress)

This is an implementation of SCINet using tensorflow and a work in progress. I want to explore the possibility of using SCINet to predict cryptocurrency prices and how they compare to traditional approaches such as an ARIMA.

SCINet is a novel architecture for time series forecasting proposed in this paper. See original paper for link to datasets.

Notes

  • See applications.testing.sinewave.py for usage examples
  • Obtained similar results on the ETD dataset (ETDataset-main/ETT-small/ETTh1.csv) used in the orignal paper but only with a batch size of 16 instead of 4. The cause of the discrepancy is unclear - pending investigation.
  • Scored poorly on crypto data (mse ~= 1.5, ase ~= 0.8 when data is relative difference). Learning curve suggests model is underfitting, which is expected as the data contains only a few basic features and has undergone minimal feature engineering. No hyperparamters tuning either. The score should serve as a baseline for future improvements.

WARNING! This is only a rough proof-of-concept and is no longer actively maintained. Use at your own risk.

Note that the project description data, including the texts, logos, images, and/or trademarks, for each open source project belongs to its rightful owner. If you wish to add or remove any projects, please contact us at [email protected].