facebook / facebook/prophet

Abnormally increased value in prediction section

Open
#2,077 1 comment 0 reactions 0 assignees View on GitHub
question
Dominant language
Python
Stars
20.4k
Forks
4.6k
Avg merge
19h 52m
Merged PRs (30d)
1

Description

Hello!

I'm using prophet to predict future sales.

In input data, the sales volume treated with squared root is used as 'y' value.
And cap and floor are each 100 and 0.
Additionally, price information is added with `add_regressor`

But some of the products , its result seems strange.
The value is increasing abnormally in the prediction section.
There is no holiday event on the prediction section.

Below is the example & input data.

### sample graph

![img](https://github.com/jiyeoon/TIL/blob/main/etc/sample_graph.png?raw=true)

### Input data

ds | y | price | cap | floor
-- | -- | -- | -- | --
2020-11-01 | 3.464102 | 970.0 | 100.0 | 0.0
2020-11-02 | 1.000000 | 970.0 | 100.0 | 0.0
2020-11-03 | 3.000000 | 970.0 | 100.0 | 0.0
2020-11-04 | 3.000000 | 970.0 | 100.0 | 0.0
2020-11-05 | 3.741657 | 970.0 | 100.0 | 0.0
... | ... | ... | ... | ...
2021-10-21 | 2.449490 | 990.0 | 100.0 | 0.0
2021-10-22 | 2.449490 | 990.0 | 100.0 | 0.0
2021-10-23 | 2.236068 | 990.0 | 100.0 | 0.0
2021-10-24 | 2.449490 | 990.0 | 100.0 | 0.0
2021-10-25 | 1.732051 | 990.0 | 100.0 | 0.0

### code

```python
model = Prophet(
growth='logistic',
holidays = holidays
)
model.add_country_holidays(country_name='KR')
model.add_seasonality(name='monthly', period=30.5, fourier_order=5)
model.add_regressor('price')
model.fit(data)
```

Why is the happening?

Contributor guide

Open the contributing guide

Assessment

This issue has not been assessed yet.

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.