Questions tagged [forecasting]
Prediction of the future events. It is a special case of [prediction], in the context of [time-series].
3,953 questions
1
vote
0
answers
37
views
Designing a demand forecasting model with a dynamic daily update and a final horizon prediction — best practices to avoid leakage?
I am working on a demand forecasting problem for ferry vehicle capacity.
For each voyage, I have daily snapshots of the cumulative reservations from the opening date until departure day.
So each ...
0
votes
0
answers
9
views
Are there any other powerful optimization tools available besides the ABC and PSO algorithms? [duplicate]
What are other optimization tools that are powerful enough to improve the accuracy performance of the neural network model? Please give me recent tools that are powerful
- 823
1
vote
0
answers
50
views
What is the best statistical approach to forecast cash flow from run-off debt vintages with a growing balance?
community.
I'm facing a modeling problem for cash flow forecasting and would like to know what the most robust mathematical/statistical approach is to solve it.
The Problem: Debt Recovery Forecasting
...
- 215
1
vote
0
answers
33
views
Need advice on length of context for future prediction
I'm using a trained foundation model to forecast values on a time series. The model works by taking a window of recent data (context) to predict near-future outcomes (horizon).
How can I know the ...
- 11
1
vote
0
answers
18
views
How to use a hierarchical Bayesian model to combine regional and country-level data for TPES projections?
I’m trying to project TPES (Total Primary Energy Supply) by country in Africa up to the year 2100 under different SSP (Shared Socioeconomic Pathways) scenarios, the same framework used in the latest ...
1
vote
1
answer
46
views
Forecasting with VAR models after differencing some time series
i have data from 85 participants who answered 6 items for 82 consecutive days. In other words: I have 6 time series per participant. I already imputed the missing data, so that all time series have ...
- 61
4
votes
1
answer
147
views
Is using contemporaneous components to forecast an aggregate a valid method or a form of data leakage?
I am in the middle of a deep methodological debate regarding a time series forecasting problem and would appreciate the community's expert opinion.
The Context
I am trying to forecast an aggregate ...
- 298
0
votes
0
answers
50
views
Sampling counterfactual posterior to mitigate error autocorrelation in event studies
I have question regarding event studies (pre-event data is observed, an event occurs at $t=e$, then following the treatment is assumed to be in-effect.)
There are multiple approaches to event study ...
- 3,970
0
votes
0
answers
40
views
Forecasting beta distributions
I hope you are doing well. I would appreciate your help with the following questions (listed at the end).
Context
There is a financial aid program that covers tuition fees for university students for ...
- 133
5
votes
4
answers
1k
views
What is difference between nowcasting and forecasting?
Recently I crossed to this Github Repo trying Benchmarking econometric using ML models in nowcasting GDP (see the paper).
Q1: What is difference between nowcasting and forecasting over time data in ...
- 579
1
vote
0
answers
45
views
Dynamic Linear Model: Classical vs Discount Approach
I'm working on a time series forecasting problem and trying to decide between the classical approach and discount approach for Dynamic Linear Models (DLMs).
Anyone here has experience comparing these ...
- 73
4
votes
1
answer
196
views
Can I have some help choosing a very low-sample estimator?
I want to forecast what next semester's finances may look like, regarding my campus job. I get paid bi-weekly, and have eight past data points: ...
- 143
-1
votes
2
answers
191
views
Quantile regression: But which quantiles (conditional on what)?
I have a task of making a quantile regression (5%, 50% and 95%) for tomorrow's power production. However, I am trying to grasp which quantiles we are talking about. Wikipedia (and similar sites) ...
5
votes
1
answer
172
views
Distribution based loss for regression with unbounded data
Currently I am dealing with time-series data conserning the power consumption of machines. Therefore, all target variables range from zero to infinity, technically ($y \in [0, \infty)$). The data ...
- 153
4
votes
1
answer
93
views
5
votes
1
answer
156
views
How to handle missing weather data (predictors) in production time series forecast with Prophet?
I'm currently running a production pipeline that uses Facebook Prophet (GAM) to forecast future electricity usage. The model includes:
Target: past electricity consumption (hourly data)
External ...
- 53
1
vote
1
answer
99
views
Estimating Box-Cox lambda, Guerrero vs maximum likelihood estimation (MLE)?
In the textbook “Forecasting: Principles and Practice (3rd ed)”, at https://otexts.com/fpp3/transformations.html , I see:
“3.1 Transformations and adjustments ...
Mathematical transformations
If the [...
- 31
0
votes
0
answers
64
views
Time series LASSO K-fold cross validation
This topic has been discussed before but I couldn't find a specific answer.
Here's my approach to forecast QoQ values,
Run the usual LASSO K-fold CV on timeseries data and generate a one-step ahead ...
- 1
1
vote
1
answer
87
views
Is it correct to use SHAP to explain actual observed values?
I have a tree-based model trained for demand forecasting and SHAP is the method chosen for explaining predictions. Among the features are history lags, promotions, pricing, resizing and many demand ...
- 231
3
votes
0
answers
185
views
Interrupted time series analysis with several interventions during the study period
I am conducting an interrupted time series analysis. The time series is the monthly incidence of new users of a given medication from January 2015 to December 2024 in France. I want to assess the ...
- 600
0
votes
0
answers
49
views
One-Step Ahead Forecasting with TensorFlow Structural Time Series
I have the following situation: I’m given a univariate time-series dataset $y$ that I wish to model using feature variables $X$, which are provided alongside $y$. Naturally, I split the data into a ...
0
votes
0
answers
100
views
Need Help with ARIMA Analysis – Stationarity and Differencing Confusion
I’m currently working on another ARIMA analysis using a different yearly dataset I obtained online. The dataset contains 42 observations with two variables: Year and Number of Deaths. I've already ...
- 31
0
votes
0
answers
98
views
ARIMA Modeling on Aggregated Global Data: Stationarity, Differencing, and Forecasting Concerns
Good day! I’m currently working on a time series analysis using ARIMA for a global yearly dataset. The data comes from a publicly available and anonymized dataset (obtained from the Global Burden of ...
- 31
3
votes
1
answer
184
views
Need Help with ARIMA Modeling on Yearly Global Data
I am currently working on my time series analysis, which I am still new to. My dataset is yearly and involves global data on selected univariate variables. I have followed the steps below, but I’m not ...
- 31
1
vote
1
answer
71
views
Seasonalize a time series with the components from other series
I have weekly sales data for thousands of products, a huge portion of them present clear seasonal patterns, weekly and monthly.
My issue is that the data was recorded weekly onle since 2023, so all ...
- 231
0
votes
2
answers
125
views
Combine conditional mean and variance equation to estimate point predictions in ARMA-GARCH model?
An ARMA(p,q)-GARCH(r,s) model specifies the conditional distribution of a time series $x_t$:
\begin{aligned}
x_t &= \mu_t + u_t, \\
\mu_t &= \varphi_0 + \varphi_1 x_{t-1} + \dots + \varphi_p ...
- 559
0
votes
0
answers
17
views
Why does my GNN-LSTM model fail to generalize with full training data for a spatiotemporal prediction task? [duplicate]
I'm working on a spatiotemporal prediction problem where I want to forecast a scalar value per spatial node over time. My data spans multiple spatial grid locations with daily observations.
Data Setup
...
- 43
5
votes
1
answer
138
views
Looking for Tips on Forecasting Seasonal Inventory Data
I run a small shop with around 500 products and, in my spare time, I have decided to create a forecasting model to help manage my inventory better. At the moment, I sometimes end up with too much ...
- 51
3
votes
0
answers
158
views
Evaluating resolution of a density forecast of a continuous variable
In the context of density forecasting, resolution addresses the following question: Is the conditional distribution different from the unconditional distribution? The concept is presented briefly in ...
- 71.5k
0
votes
0
answers
33
views
Forecasting with incomplete data
I'm looking for some guidance on approaches to forecasting where the most recent data is not complete.
I work for a health insurance company where the data takes months to reach steady state i.e. ...
- 649
2
votes
0
answers
59
views
Best model to combine predictors [closed]
I have a few curves that predict the same outcomes, all curves are extremely similar but vary a little in terms of noise and predictions (guessing they have lots of similar variables and some ...
- 21
0
votes
0
answers
58
views
Seeking advice on time series predictive modeling from experts
I am looking to identify outliers (faulty sensor data) in hourly DHW-demand (domestic hot water) from 4 years.
And I have 4 years of hourly weather data that is pretty much accurate and a majority of ...
- 1
1
vote
0
answers
53
views
How do I go about refining my ARX model in R
I face a few issues where im trying to predict my dependent variable Y. I have 6 different independent external variables with one of them being lag(1) of the dependent variable Y. I differenced all ...
- 11
2
votes
0
answers
77
views
Uncertainty in sum of K new samples from an estimated distribution
I have $N$ independent samples from an unknown normal distribution $\mathcal{N}$. If I draw a single new sample from $\mathcal{N}$, I know from literature that a confidence interval for this new ...
0
votes
1
answer
104
views
Data Leakage with STL Decomposition
I have a time series that I need to forecast with a SARIMA model. I do a train test split and then fit the SARIMA model on just the training data. I want to avoid data leakage and preserve realism, so ...
- 143
1
vote
0
answers
76
views
Why minimise Calibration Error rather than MSE? Context: LLM Hallucination [closed]
In the discussion of Large Language Model hallucination phenomenon, people are interested in measuring and reducing the calibration error of the model predictions. However, what makes this situation ...
- 111
0
votes
1
answer
91
views
Applying the Diebold-Mariano Test with a Decreasing Forecast Horizon
I’m using the Diebold-Mariano test to compare the forecast accuracy of two forecasts, but my dataset has a unique characteristic: the forecast horizon decreases over time. As the event date approaches,...
- 3
3
votes
0
answers
102
views
In density forecasting, is there a standard measure of sharpness?
In the context of density forecasting, Gneiting et al. (2007) characterize sharpness as follows:
Sharpness refers to the concentration of the predictive distributions and is a property of the ...
- 71.5k
2
votes
1
answer
105
views
Why is negative probability density an improper scoring rule? A counterexample?
Gneiting and Katzfuss (2014) discuss (among other things) proper scoring rules for evaluating density forecasts. Quoting the paper (p. 133),
Definition 4: The scoring rule $S: F \times R \rightarrow \...
- 71.5k
7
votes
3
answers
389
views
Adjusting Brier score for the "easiness of a bet"
I'm working on a project that depends on evaluating forecasting ability of users of Manifold, which is a site where you can bet on events and earn play money. Since there is an application that allows ...
- 301
0
votes
0
answers
29
views
smoothing parameters
I am doing a cases study and I need to forecast the sales. I am using multi linear regression and also winter's method and the decomposition approach with Holt's method. I am using these methods as ...
- 1
2
votes
0
answers
55
views
Python based forecasting textbook
I am on the hunt for a good textbook on business forecasting for graduate students with examples in Python. I have not been able to find any yet. I have taught using the "Hyndman and ...
1
vote
0
answers
76
views
Using Pre-whitening as a Variable Screening Method for Time Series Forecasting
I am working with a large pool of economic variables (similar to FRED-MD dataset) with the following characteristics:
Number of observations (n): 250
Number of predictor variables (p): ~5,000
Ratio p/...
- 11
1
vote
0
answers
123
views
How can I modify my binary decision function to capture sparse 1’s without averaging out the signal?
I’m forecasting whether a specific customer will purchase a specific product in each of the next 12 months (1 if purchased, 0 if not). My historical monthly data for this customer/product is extremely ...
4
votes
1
answer
130
views
Promotional effect on demand
I have 3 years of weekly sales data for several product categories. I am interested in analysing the immediate and lagged effects of price promotions. The price promotions are limited to a few weeks ...
- 41
0
votes
0
answers
24
views
Modeling Time series with ARIMA [duplicate]
I am trying to model a time series. I am not certain if my approach is correct and that's why I am posting a question here. May be someone could point out the ...
- 161
5
votes
1
answer
495
views
Time series models - ML based vs classical methods - cases with changing trends and additional features
I'm new to the field of time series and have been reading up about it. It seems there's no overwhelming consensus in favor of either ML or classical methods for forecasting problems. ML methods seem ...
- 253
4
votes
1
answer
365
views
What is the name of $\min(a,f)/\max(a,f)$ as an alternative to SMAPE?
A coworker of mine asked for an accuracy measure. In the past I was taught to use absolute percent error (ape) ... well actually the compliment (1 - ape). This way bigger numbers are good.
This works ...
- 2,093
4
votes
1
answer
77
views
How can I test whether a forecast used another forecast as input?
We're running a forecasting competition on renewable energy generation. A forecast $\hat Y_{ref}$ of the target variable ($X$) is already published (by us, actually. But we can't touch it). ...
- 211
6
votes
1
answer
311
views
Which model/technique to use to forecast seasonal revenue peak?
I have 1000s of customers and their monthly revenue data (in €) for the last 10 years. Some display seasonal/cyclical patterns, others display ...
- 161