Comments (3)
And
- how to merge batch generated sequences(segmented data)? Simply concatenate may introduce unreasonable temporal dynamics between each adjacent segment pair.
Thanks again.
from timegan.
Thank you for your interest in our paper.
- Static features
- As you know, static features are defined as the variables whose values do not depend on time. (Time-invariant variables) such as gender, race, etc.
- Unfortunately, in Stock data, there is no static features.
- Therefore, we do not implement the static feature parts in the paper in this GitHub implementation.
- Stationary vs Static
- I think stationary and static is different.
- Random process can be a stationary process but the static mean in the paper is not about the stationary process.
- Merging
- In the experiments, we assume that all the data comes from iid distribution.
- So, when we cut the time-series data (Stock data) and mixed (permuted), we already treat them as iid.
- Then, the focus of TimeGAN is to generate those sliced time-series and not the original time-series.
- Therefore, there is no way to merge it.
- If you want to generate the original stock data, I think you need to use different preprocessing procedures.
from timegan.
Thanks for your detailed reply. I'll do stationary transform in preprocessing.
:)
from timegan.
Related Issues (20)
- Sigmoid activation function in recovery (decoder) network HOT 1
- module 'tensorflow' has no attribute 'reset_default_graph' HOT 2
- Several questions regarding architectural choices HOT 1
- Wrapper function to generate new data? HOT 1
- How to search for optimal hyperparameters with TimeGAN? HOT 1
- How to search for optimal hyperparameters with TimeGAN (continued)? HOT 1
- Supervised Loss HOT 2
- version of python HOT 1
- About the form of the generated data HOT 6
- About the discriminator HOT 1
- A question about evaluating the predictive score HOT 1
- Question about reconstruction of generated sample HOT 1
- How to apply 1-dimensional data to the timegan framework HOT 4
- Reducing dimensionality with np.mean() HOT 2
- On "Mix the datasets (to make it similar to i.i.d)" HOT 1
- Saving Model During Training and Using Generator Independently of Training HOT 1
- `G_loss_S` does not depend on the generator variables `g_vars`, no need to add it to the solver HOT 1
- The usage of 2 embedder losses (`E_loss_T0` and `E_loss`) HOT 1
- timeGAN 1D HOT 2
- Flipping the data in dataloading.py HOT 1
Recommend Projects
-
React
A declarative, efficient, and flexible JavaScript library for building user interfaces.
-
Vue.js
🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
-
Typescript
TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
-
TensorFlow
An Open Source Machine Learning Framework for Everyone
-
Django
The Web framework for perfectionists with deadlines.
-
Laravel
A PHP framework for web artisans
-
D3
Bring data to life with SVG, Canvas and HTML. 📊📈🎉
-
Recommend Topics
-
javascript
JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
-
web
Some thing interesting about web. New door for the world.
-
server
A server is a program made to process requests and deliver data to clients.
-
Machine learning
Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
-
Visualization
Some thing interesting about visualization, use data art
-
Game
Some thing interesting about game, make everyone happy.
Recommend Org
-
Facebook
We are working to build community through open source technology. NB: members must have two-factor auth.
-
Microsoft
Open source projects and samples from Microsoft.
-
Google
Google ❤️ Open Source for everyone.
-
Alibaba
Alibaba Open Source for everyone
-
D3
Data-Driven Documents codes.
-
Tencent
China tencent open source team.
from timegan.