Comments (8)
hey, we're looking into this case now. it does not always occur and thus bit difficult to debug for us. we'll report back asap. also, please make sure the set_hyperBoundLocal is set correctly (right names using local_name, right bounds according to guide) and set_mediaVarName and set_mediaSpendName are set correctly
from robyn.
Thanks @gufengzhou .
Further context if it can help:
- I have a 1to1 mapping between set_mediaVarName and set_mediaSpendName (aspend metric for each pressure metric)
- adstock = 'geometric'
from robyn.
I think we found the root cause. What train/test split are you using/ what value do you put in set_modTrainSize? One extreme example, you have 2 years data and one media channel didnt have any spend until 2 recent months. When you set set_modTrainSize = 0.8, that media channel will only contain 0 in the training set and that causes problem in the transformation. What you can already do is set appropriate training split by looking at the plot f.plotTrainSize(T), the higher the bhatta coef the better. Basically you should avoid any channel only containing 0 in the train set. I'll push a fix to warn this condition
from robyn.
Yep. That's likely to be the issue. I have channel that were switched on only recently.
I'll play around with f.plotTrainSize(T) to find the best bhatta coef and comeback to you
Current train-test split
set_modTrainSize <- 0.74
from robyn.
I got this output with `f.plotTrainSize(T)'
I have channels with data only for the last couple of months so I'll work backwords by trying a model with data for most of the weeks (rows)
from robyn.
Issue resolved by aggregating/remove channels containing 0 in the training set.
Thank you very much
from robyn.
FYI, I've pushed an update to deal with/better warn about situation where some media has many 0. Please use the new version. thanks again for the great use case.
from robyn.
thanks. Updated and run successfully locally 🙌
from robyn.
Related Issues (20)
- demo.r | document error HOT 1
- robyn_allocator - What's the difference between expSpendUnitTot & optmSpendUnitTotal? HOT 2
- Robyn_outputs: vector size error HOT 2
- Robyn Allocator underestimates Carryover HOT 23
- robyn_recreate() error: Must provide 'hyperparameters' in robyn_inputs()'s output first HOT 3
- Robyn CSV Data Dictionary? HOT 2
- Does Robyn still use only spend data as input but not exposure metrics? HOT 4
- Attribution for Brand Direct HOT 2
- Error in {: Task 1 failed - "object 'dt_geometric' not found" HOT 3
- Impression modelling HOT 4
- Total response in robyn_allocator output HOT 1
- Upper Boundary error in Budget Allocator HOT 9
- Refresh output is not matching the documentation HOT 3
- Output of robyn_recreate() returns warning message referencing an object that does not exist anymore? HOT 1
- Different Results in Recreated Model HOT 7
- Error in NRMSE calculation HOT 1
- Import mode output to run optimizations HOT 2
- Error in robyn_outputs step when recreating model from previous model inputs and hyperparameters HOT 5
- Factorial Variables HOT 1
- Budget Allocator Question HOT 9
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 robyn.