Comments (6)
Conceptually we can handle multiple observations on a day as coming from the noise term, but could you give some context for the application that you have in mind?
from prophet.
The system I'm working on captures transactions end of day and the timestamp is truncated to just the date. So, if there are more than one, I'd typically just take the mean and go with it.
The thing is, Prophet's error message is not easily interpreted as this problem. It just says,
TypeError: int() argument must be a string, a bytes-like object or a number, not 'slice'
Perhaps, a more user-friendly error message could help to begin with..
from prophet.
Thank you for capturing the duplicate error @xhesc. I have a similar use case. I am projecting ticket sales, but on some evenings we have more than one show, and thus more than one entry for a given date. In this case a mean doesn't make as much sense and grouping by date (day, week, month) would make more sense.
from prophet.
The latest commit should fix the issue with not allowing multiple observations on a date. Could you install the Github version and check if it works for you?
The model now treats multiple observations on a day similar to how a linear regression would. We do assume homoscedastic noise, which here means that the variance in the values for a single day is constant. @korymath I can imagine in your application there would be a difference in variance from weekdays vs. weekends, but if you sum sales in a day then I'd think weekly seasonality should capture that nicely.
Now that repeated observations aren't a problem, I'm inclined to leave other aggregations as a data pre-processing step that would be taken care of outside of Prophet.
from prophet.
I think you are right. Nice job with the fix and I agree that data pre-processing can help other cases. Great work.
from prophet.
Looks good python wise, didn't test R.
Great work on commit: 443d475
from prophet.
Related Issues (20)
- Time zone issue when fitting prophet in R HOT 1
- Minimum Supported Python version is 3.8, not 3.7
- Fitting hangs in Prophet 1.1.5
- weired x label in model.plot_components(forecast)
- ries
- Hierarchical Model
- Model fit behaves very differently between `datetime64[ns]` and `datetime64[ms]`.
- Amplifying the amplitude of seasonality
- regressor_coefficients not finding regr_names HOT 1
- sample_model() not loading
- Forecast is off by 1
- Prophet Optimization error HOT 1
- Can't seem to forecast values correctly for future dates (few values are coming zero and few are not matching the historical pattern)
- Facebook Prophet ignores manually passed changepoints
- Prophet not working on Google Cloud Run HOT 1
- While adding regressor receiving error message: TypeError: ufunc 'isnan' not supported for the input types HOT 1
- Prophet not compatible with Pandas 2.0
- If there is no data in holidays
- How to get rid of the time dependance?
- Optimizing the training result
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 prophet.