patrick-kidger/Deep-Signature-Transforms
Code for "Deep Signature Transforms" (NeurIPS 2019)
This project offers a novel way to build neural networks for processing sequential data. It takes raw data streams, like financial time series or sensor readings, and transforms them into a compact set of features (a 'signature') that effectively summarize the stream's characteristics. This is then fed into a neural network for tasks like prediction or classification. Data scientists and machine learning engineers working with complex time-series data will find this useful.
No commits in the last 6 months.
Use this if you are a machine learning engineer or researcher designing neural networks for time-series analysis and want to incorporate a powerful, mathematically rigorous feature extraction method directly into your model's learning process.
Not ideal if you are looking for a pre-trained, out-of-the-box solution for general time-series forecasting without needing to delve into neural network architecture design.
Stars
98
Forks
20
Language
Jupyter Notebook
License
Apache-2.0
Category
Last pushed
Jul 25, 2024
Commits (30d)
0
Get this data via API
curl "https://pt-edge.onrender.com/api/v1/quality/ml-frameworks/patrick-kidger/Deep-Signature-Transforms"
Open to everyone — 100 requests/day, no key needed. Get a free key for 1,000/day.
Higher-rated alternatives
lululxvi/deepxde
A library for scientific machine learning and physics-informed learning
pnnl/neuromancer
Pytorch-based framework for solving parametric constrained optimization problems,...
wilsonrljr/sysidentpy
A Python Package For System Identification Using NARMAX Models
dynamicslab/pysindy
A package for the sparse identification of nonlinear dynamical systems from data
google-research/torchsde
Differentiable SDE solvers with GPU support and efficient sensitivity analysis.