Adding data having same TimeStamp but different source

[abhishekju06]

Issue description

Hi Wenjie,

I want to prepare my dataset having same Timestamp but from different source. Can you please let me know how to do it?

[WenjieDu]

Please don't ask me too-general questions like this, much less it's not related to PyPOTS. I'm not a free tutor. If you want to raise a discussion, please at least show that you gave your thought to it. Google it and do your research first. Thanks.

[abhishekju06]

I didn't find any such preprocessing code that supports the generation of dataset having same timestamp but different sources in PyPOTS. That was the reason I asked you. Thanks for such reply.

[WenjieDu]

Out of question, a library couldn't and also is impossible to contain all functions that users may need, especially for such a personalized step as data preprocessing.

Some suggestions here. You can try to merge all features together, say you have 3 features for data from one source and you have 5 sources, then you have 15 features after merging. This is like the preprocessing of the dataset AirQuality in SAITS paper. Or you can organize different sources into one more dimension, say it location dim. Then you need models having the capability of handling the data with such one more dimension, like spatial-temporal models.

[abhishekju06]

Hi Wenjie, Hope you are doing great. Just wanted to know how to load and predict a model that I have trained earlier. The models are saved in folder ..\saits\20230718_T203937. I cannot figure this out. Please help

…

On Mon, Jun 19, 2023 at 12:24 PM Wenjie Du ***@***.***> wrote: Out of question, a library couldn't and also is impossible to contain all functions that users may need, especially for such a personalized step as data preprocessing. Some suggestions here. You can try to merge all features together, say you have 3 features for data from one source and you have 5 sources, then you have 15 features after merging. This is like the preprocessing of the dataset AirQuality in SAITS paper. Or you can organize different sources into one more dimension, say it location dim. Then you need models having the capability of handling the data with such one more dimension, like spatial-temporal models. — Reply to this email directly, view it on GitHub <#146 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AJLYKN7LQFYJMTFU2XQHK3TXL7ZS5ANCNFSM6AAAAAAZLNGJIE> . You are receiving this because you authored the thread.Message ID: ***@***.***>

[abhishekju06]

Sorry to bother you. I was able to figure it out with saits.load_model(model_path). On Fri, Jul 21, 2023 at 12:38 PM Abhishek Das ***@***.***> wrote:

…

Hi Wenjie, Hope you are doing great. Just wanted to know how to load and predict a model that I have trained earlier. The models are saved in folder ..\saits\20230718_T203937. I cannot figure this out. Please help On Mon, Jun 19, 2023 at 12:24 PM Wenjie Du ***@***.***> wrote: > Out of question, a library couldn't and also is impossible to contain all > functions that users may need, especially for such a personalized step as > data preprocessing. > > Some suggestions here. You can try to merge all features together, say > you have 3 features for data from one source and you have 5 sources, then > you have 15 features after merging. This is like the preprocessing of the > dataset AirQuality in SAITS paper. Or you can organize different sources > into one more dimension, say it location dim. Then you need models having > the capability of handling the data with such one more dimension, like > spatial-temporal models. > > — > Reply to this email directly, view it on GitHub > <#146 (comment)>, > or unsubscribe > <https://github.com/notifications/unsubscribe-auth/AJLYKN7LQFYJMTFU2XQHK3TXL7ZS5ANCNFSM6AAAAAAZLNGJIE> > . > You are receiving this because you authored the thread.Message ID: > ***@***.***> >

[WenjieDu]

No problem. You can follow me to receive the latest news about PyPOTS.