Questions tagged [prophet]

An automatic forecasting system developed by Facebook. Use this tag for any on-topic question that (a) involves Prophet either as a critical part of the question or expected answer and (b) is not just about how to use Prophet.

Filter by
Sorted by
Tagged with
12
votes
3answers
809 views

Why is non-iid noise so important to traditional time-series approaches?

I've been reading the whitepaper that accompanied Facebook's release of Prophet, it's time-series modeling library. One topic the authors drew attention to was that noise was assumed to be iid; they ...
0
votes
0answers
6 views

Using historical proportions for prediction intervals estimated at a more aggregated level?

Currently, I am working on quite a general problem where I need to forecast some business-related value like target events (clicks, acquisitions, etc). Given the number of different breakdowns (by ...
1
vote
0answers
10 views

cross validation and hyperparameter tuning for multiple time series

I have time series data for 2000 products. If I use models like fbprophet or SARIMAX or xgboost then the cross validation needs to be done for 2000 time series data. for a single time series data it ...
0
votes
0answers
26 views

Is a Laplace Prior the same thing or related to a Laplace Transformation?

Context: I was watching this video https://youtu.be/pOYAXv15r3A?t=796 about Facebook Prophet and the speaker mentioned they use a Laplace Prior $$\delta \sim Laplace(\lambda)$$. What I have gleaned so ...
0
votes
0answers
54 views

Forecasting a distribution

My current struggle is to forecast the distribution of a given value across the next 360 days. e.g. The context is hotel accomodation bookings. We have a forecast of 100 bookings to be made for time t ...
0
votes
0answers
14 views

Selecting values for time series cross-validation (Facebook Prophet)

Let's say I have a problem with ~400 daily obervations and I need to forecast 7 days ahead. The goal is to report how accurate such forecasts can be. Facebook Prophet provides a convenient method <...
0
votes
1answer
43 views

Forecasting problem - Fewer data points and covid impact

I have quarterly data from 2017 to 2020 (16 data points) to forecast. I tried to use SARIMA but it is giving me weird numbers (High and negative fitted values). Also, I tried exponential smoothening ...
2
votes
0answers
32 views

How to transform a daily average temperatures forecast into an hourly forecast based on the hourly temperature profile observed historically?

I need to transform a daily average temperatures forecast into an hourly forecast based on the hourly temperature profile observed historically. I work in Python. I have found ways of forecasting the ...
0
votes
0answers
9 views

Components plots is inconsistent with df_pred on monthly seasonality

Background I do seasonality decomposation with timeseries. In my use case, I have trend/weekly(7 days)/monthly(30.5 days as prophet doc suggests) components. For me, I want to get the data of each ...
2
votes
0answers
81 views

What is the best approach to grouped time-series forecasting?

Let's say we have data on the number of clicks per user over quite a long period of time. We can use, say Facebook Prophet, to forecast daily values given that we have enough historical data. That ...
0
votes
0answers
27 views

Forecasting Sales using Prophet in R

Could someone please help me figure out how to adjust my code so that it gives better predictions? Right now they are way to high. I'm wondering if it is something to do with the holiday lag. I'm not ...
0
votes
0answers
43 views

Understanding Facebook's Prophet

I am having trouble with understanding facebook's prophet's cross validation. I have an ARIMA model that splits the data into training and test sets and performs a rolling forecast by using the train ...
0
votes
0answers
98 views

Unsupervised anomaly detection on multivariate time series

Me and my team face the next use case: Our data consists of 3 numerical signals, Which are collected every 10 minutes. Example: Our main goal is to build an anomaly detection algorithm. Our work till ...
0
votes
0answers
278 views

What is the difference between sliding, rolling and expanding window in Time series forecasts?

What exactly is the difference between sliding, rolling and expanding windows in time series forecasting? Are rolling and expanding windows just subsets of sliding window? And which one of those ...
0
votes
0answers
20 views

doing a cross validation mean squared error on a time series data using fable's prophet

I'm trying to do a time series cross validation with the fable prophet model on an increasing window for training and a h = 1 lag for estimation. Here's my code ...
2
votes
1answer
51 views

Adjustment factor in logistic growth model of facebook-prophet

In the paper it emphasizes: "When the rate k is adjusted, the offset parameter m must also be adjusted to connect the endpoints of the segments. The correct adjustment at changepoint j is easily ...
1
vote
0answers
272 views

Why does prophet produce much tighter prediction intervals than ETS?

I'm currently working on a forecast problem, where narrow prediction intervals are preferred. When I look at the prediction intervals of ETS and prophet forecasts, I'm surprised that the prophet ...
1
vote
1answer
286 views

What could cause facebook's Prophet model to do so poorly on these procedurally generated datasets, where one is a continuation of the other [closed]

Recently I've been looking into some easy out of the box modeling using Facebook's Prophet -- potentially to use in some projects at work. So far, I have been super impressed with everything that I've ...
1
vote
1answer
261 views

Can you combine Facebook Prophet (fbprophet) with ARMA?

I am new to dealing with time-series data, so my apologies if this is not a valid question. I am wondering if there is a way to combine ARMA with fbprophet. From what I understand, fbprophet can take ...
2
votes
0answers
225 views

What are the underlying statistical differences between the vector autoregression model and Prophet?

I am trying to understand the underlying fundamental/statistical differences between vector autoregression models and Facebook's Prophet, with regards to multivariate time series forecasting. I am ...
0
votes
0answers
47 views

How to predict unknown time series in using Facebook Prophet?

The problem is predicting hits on different stories published on a website. I am aware that for this kind of time series forecasting, facebook prophet is a popular library. However, it seems in ...
1
vote
1answer
817 views

Why do time series, with positive values decomposed into seasonal plots, have negative values

I have been running different forecasting algorithms such as Facebook Prophet and Forecast (from R) and I note that despite all my time series values being positive, my seasonality values are negative....
5
votes
2answers
2k views

Inference in Time Series: Prophet vs. ARIMA

I read through Prophet's white paper and they mention that their algorithm, "gives up some important inferential advantages of a generative model such as an ARIMA." (page 7) So now I'm ...
2
votes
1answer
118 views

Is summing daily forecasts a sound method for generating weekly/monthly forecasts?

I'm new to time series analysis, and I am wondering if this is a sound method for generating weekly and monthly predictions. In my case, I need to generate daily, weekly, and monthly predictions. If ...
0
votes
1answer
66 views

logistic growth and covid cases/deaths

A few blogs suggest that at least the growth of covid cases (and deaths?) follows a logistic groth curve. See: https://medium.com/katanaml/covid-19-growth-modeling-and-forecasting-with-prophet-...
1
vote
1answer
3k views

Why yhat fbprophet returns with negative prediction

I'm trying to forecast with fbprophet, the input are all positive but the predictions returns negative i'm kind of confused, i read this quick start and if the inputs are all positive then the ...
0
votes
1answer
103 views

prophet cross validation

I'm confused by the 'cross_validation' of prophet. In the following cross validation process, were parameters learned and saved to the model? is this cross validation used to train model or just to ...
1
vote
0answers
44 views

how to calculate the confidence interval when sum the predictions

I want to predict the total sales of 10 stores. 8 of the 10 stores have 10 years sales data available, while the other 2 stores only have 2 and 4 years sales data. Firstly, I use prophet to build one ...
0
votes
0answers
52 views

R - How to create a forecast object from a pre-defined forecast, in order to apply accuracy() for MASE result

I am working with a forecast which has been created by Prophet. I would like to apply accuracy() to return the MASE after identifying this as the best accuracy measure for multiple forecasts, and will ...
1
vote
0answers
184 views

why prophet doesn't capture some holiday effect well [closed]

I used prophet for time series forecasting. the holiday effect for Christmas and thanksgiving were captured well. The data show strong positive effect for the Easter Sunday('lower_window': -2, '...
1
vote
1answer
104 views

prophet holiday effect

I use prophet to do time series forecasting. During Christmas season, the data rise from 12/22, reach peak on 12/23 and 12/24 (positive effect), then drop a lot on 12/25 and 12/26 (negative effect). ...
0
votes
0answers
352 views

I am trying to build a monthly revenue prediction model using Prophet in Python but how can I optimize the hyperparameters of the model?

The hyperparameters which I am trying to optimize are: 'n_changepoints', 'changepoint_range', 'holidays_prior_scale', 'seasonality_prior_scale' and 'changepoint_prior_scale'. 'changepoint_range' has a ...
1
vote
0answers
65 views

How do additional regressors and seasonality get incorporated into fbprophet? [closed]

I've been playing around with fbprophet here and I've noticed they specify their base additive model is $$y(t) = g(t) + s(t) + h(t) + \epsilon_t$$ My question is two-fold: 1) How does modifying ...
1
vote
0answers
706 views

Dealing with non continuous data in time series

My current project involves building forecasting model for asset - desktop, laptop and monitors. The data is for the last 3 years. There is good amount of seasonality in the data. There are few data ...
0
votes
1answer
55 views

Data aggregation with facebook prophet

I'm using facebook's prophet library for forecast, the library seems to work very well but I'm not finding in the documentation how to handle multiple values for the same period, E.G: I'm working ...
3
votes
1answer
198 views

Standardise components of an additive model output

I've got a sales forecasting model using the fbprophet library. The model is additive: calculates a base trend and then adds ...
2
votes
1answer
306 views

Estimate the time series like an event was never happened

I have data from a website where a specific advertising campaign happened a couple of years ago. What I want to do is to estimate how the signups on that website would have been without that big ...
2
votes
3answers
838 views

Forecasting daily data with zeros in Python

I'm currently testing some forecasts on daily sales quantities. However, out of ~2000 observations I have 16 zeros. How should I approach this? It's mainly Sundays and holidays that holds zero as ...
1
vote
1answer
667 views

Forecasting recurring orders for an online subscription business using Facebook Prophet and R

I am analyzing data from a subscription model, in which a customer must pay a recurring price at a regular interval (30 days) for access to the product. EDIT -> Direct link to daily data: https://...
15
votes
4answers
5k views

Is Prophet from Facebook any different from a linear regression?

So what I've read about Facebook's prophet is that it basically breaks down the time series into trend and seasonality. For example, an additive model would be written as: $$ y(t) = g(t) + s(t) + h(t)...
7
votes
1answer
4k views

How does facebook prophet handle missing data?

The Prophet paper (forecasting at scale by SJ Taylor - 2017) says the following on missing data: Unlike ARIMA models, the measurements do not need to be regularly spaced, and we do not need to ...