Questions tagged [forecasting]
Prediction of the future events. It is a special case of [prediction], in the context of [time-series].
3,845
questions
3
votes
2
answers
71
views
Outlier detection methods aware of target variable
I am trying to predict ambulance demand for the next hour, for a city area in the USA, based on previous demand, weather, large people gatherings, and similar spatio-temporal factors.
I have noticed ...
3
votes
0
answers
42
views
ARIMA Forecast is 14 Orders of Magnitude Higher than Training Data?
I am dealing with intermittent time series data, i.e. mostly zeros. Here is the particular time series that is giving me trouble:
[0.0,
0.0,
0.0,
0.0,
0.0,
0.0,
0.0,
60.0,
0.0,
0.0,
0.0,
0.0,
36.0,
0....
4
votes
1
answer
72
views
Manually compute ARIMAX forecast
I need to forecast a phenomenon using an autoregressive model with an exogenous variable. I've estimated an ARIMA(1,0,0) model, but I can't understand how the forecasts are calculated.
Below is the ...
3
votes
1
answer
30
views
Multiple predictions using ARIMA
I want to test the prediction capabilities of an sarima model for long sequences. I want to predict the next [24 48 96] datapoints and calculate the mse and rmse. Can you help me find bibliography and ...
0
votes
1
answer
27
views
Does moving average smoothing affect future forecasts of a time series?
I'm trying to make a multivariate time series forecast using endogenous variables. My features show a lot of spikes (noise), and as I was researching some steps to handle it, I found moving average ...
1
vote
0
answers
22
views
How to forecast changepoints from Gas Concentration Data?
So I'm trying to predict when gas concentrations change from sensor conductivity readings over a day. The gases randomly change concentrations around every 80-120 seconds and are kept constant between ...
1
vote
1
answer
32
views
SARIMAX.predict() and SARIMAX.forecast() exog? Does exog need to be preknown for predict()?
On SARIMAX.predict, when you have an exog but the exog is only known today and in the past, how do you predict the endog's next 12 months off just the exog and data ...
0
votes
0
answers
22
views
Signal detection in the presence of both theoretical and experimental uncertainty
I am wondering what the best way to go about the following situation is. I will try to frame it in a clear way but please leave a comment if something can be clarified. I suspect this type of thing is ...
2
votes
1
answer
30
views
How is forecasting values for stationary time series even possible?
Forecasting future values for a time series using Holt-Winter's model makes a lot of sense: by identifying a trend and some seasonal patterns which are likely to repeat in the future one can make ...
0
votes
1
answer
26
views
Forecasted value of ARMA model with order (p=1, d=0, q=1) scaled down and bias [duplicate]
I trained an ARMA model using Python's from statsmodels.tsa.arima.model import ARIMA .
Separated the training and testing data, and fitted the model with parameters ...
1
vote
1
answer
60
views
Predicting quantiy sold using Time series data
I am struggling with a time series dataset comprising 12 features, including quantity sold and weather data, totaling approximately 1800 values, where data is recorded on a daily basis. My goal has ...
2
votes
1
answer
23
views
Any numeric measure to indicate seasonality, uncertainty of demand?
Am working on inventory management, demand forecasting etc. As part of this project, I am exploring the data to see the past demand of the products to predict the future.
While am currently plotting ...
1
vote
0
answers
10
views
How to deal with different orders of integration between explained and explaining variables?
Is there a standard, or at least a valid, regression approach if you are trying to regress a dependent variable with a unit root against a set of stationary independent variables? I know I could ...
0
votes
1
answer
49
views
Event prediction using Statistical or Black box approach
I just came across a problem related to Event prediction in timeseries data.
So there is a timeseries data having timestamp and event occurred at that time, and I need to predict the next set of ...
2
votes
0
answers
23
views
what's the model/data drift metric for time series forecasting?
Suppose that I have time series forecasting model, e.g. forecasting point of sales revenue in various economic scenarios represented by indicators such as inflation or interest rates. I build a model $...
0
votes
0
answers
21
views
Forecasting and numerical stability - how to handle added constants post forecast?
The Python statsforecast package has some recommendations when simulating / generating time series, that we can add a constant to avoid computational problems: ...
1
vote
0
answers
24
views
Multi-step ahead forecasting using ML models with horizon feature
All academic literature I could find on this topic distinguishes between recursive forecasting (forecasting y_{t+1}, then using it as an input to forecast y_{t+2} etc) or direct forecasting using a ...
0
votes
0
answers
18
views
When do we fail to understand forecast errors and other error metrics ? How to properly use forecast error metrics to assess quality?
Whenever a forecast is calculated, it's only natural and often necessary to evaluate how good the forecast is. To evaluate the quality of a forecast of horizon $h$, we might employ a variety of error ...
0
votes
0
answers
33
views
Time Series Prediction with Partial Future Data (Presales)
I have a data set that contains tickets sold up to a certain point in time, and then presold tickets for future events. Just for demonstrative purposes, I'll use the Australian total wine sales data ...
2
votes
1
answer
66
views
Linear Regression vs. SARIMAX with Exogenous Variable: Coefficient Interpretation
I have data concerning, say, ice cream sales and wish to predict future sales and also, to quantify the relationship between temperature and sales. The data has daily seasonality.
...
0
votes
0
answers
27
views
A small part of my data is non-stationary, how to solve the issue?
For a project, I have to produce an ARIMA forecast of Boeing's stock returns over the last 10 years. I've tried to check stationarity of data with ACF first and found out that lag 1 is statistically ...
0
votes
0
answers
6
views
How should I test whether simultaneous residuals from two models of the same time series are independent?
Suppose I have two different models, with comparable goodness of fit but very different structure, that I have fit to the same time series. For both, the residuals pass various tests of normality. How ...
1
vote
1
answer
50
views
ETS (error, trend, seasonal) formulation
Does someone know if there is (clever) way to formulate mathematically all the following models below:
in a unique (system) of equations?
0
votes
0
answers
23
views
Autocorrelation of residuals in my VAR model
I am running a VAR model to predict the flow of consumer loans (dependent variable). I have three independent variables (consumption of durables goods, employment rate and households GFCF). Each ...
2
votes
0
answers
42
views
Common way to forecast multivariate time series where components are restricted by an inequality?
Let's say we have a multivariate time series that we would like to forecast future values of: $Z_t = (X_t, Y_t)$ where $X_t$ and $Y_t$ are real-valued time series and constrained by the inequality $...
0
votes
0
answers
59
views
Local linear trend-exponential smoothing duality with non-Gaussian likelihood?
Say we have the following state space/structural time series model:
$$
y_t \mid \mu, \sigma \sim \text{Normal(} \mu = \mu_t, \sigma^2_\varepsilon) \\
\mu_{t+1} = \mu_t + \eta_t
$$
where
$$
\eta_t \sim ...
0
votes
1
answer
46
views
In sliding window regression, what is the best way to select my training window and test set size?
I am trying to forecast an index option's implied volatility using a sliding window regression and I'm a little confused on how I can go about cross validating with respect to the training and test ...
0
votes
0
answers
21
views
How can I use functional time series or functional principal component regression for forecasting this type of data?
Lets say I have this type of data. For one of the series, I observed data until day 300. Can I use functional principal component regression to forecast next 50 observations? What is the best way to ...
0
votes
0
answers
5
views
Time based cross-validation for EOM patterns
I want to evaluate different models (ARIMA, MLP, LSTM, regression, etc.) on their performance to predict/forecast stock prices in a period (horizon) of 7 days around month-end. The data for these ...
0
votes
0
answers
38
views
Forecasting in ARIMA (python)
I have a time-series data (Date and Total). This is actual data from the past 2 years.
I understand how to pick the (p, q, d) order for ARIMA. And I can divide my data into train and test, and the ...
1
vote
0
answers
25
views
How can I model exponential sawtooth like regression? [duplicate]
I have this type of data. It is the occupancy rate of a hotel in January. Each line is one of the years (2019-2023). So in January occupancy rate is lets say X rooms. I know that in December it was ...
2
votes
1
answer
53
views
Sequential prediction/modeling without stationarity assumptions
Let's say that we have an arbitrary real-valued sequence of length $n$:
$$(x_{i})_{i \in \{0, \dots, n-1\}}.$$
If we wanted to try to create a probabilistic model for future values of the sequence ...
0
votes
0
answers
26
views
Forecasting: choosing the sample split between "in-sample" and "out-of-sample" data
Goals:
Given approximately 11 years of time series data, to determine how much of this data should be reserved for in-sample and ...
0
votes
0
answers
27
views
regression model with arma errors: forecasting the residuals
Suppose I estimate the following model:
$$
y_t = \beta_0 + \beta_1 x_t + \eta_t
$$
where $\eta_t$ is an AR(1) model, say. I can do that with forecast::Arima() as ...
0
votes
0
answers
21
views
forecast model has shifted result (lag = -1)
I have a problem with the LSTM forecast model; it predicts the current value instead of the next value. I have checked the data alignment, and it is correct. Try with:
different scallers and unscale ...
1
vote
0
answers
28
views
Enhancing Short-Term Sensitivity in Daily-Level Forecasting with Facebook Prophet
I have a daily time series dataset, and I'm using Facebook Prophet for daily-level predictions. Frequently, my actual data experiences sudden spikes due to external events. What I want to achieve is ...
1
vote
1
answer
50
views
in-sample and out of sample forecasting plot look very different
why the in-sample looks very different from the out-of-sample? Is the in-sample overfitting? Horizon is 12, using pycaret time series function.
0
votes
1
answer
102
views
Detecting and Forecasting Intermittent Time Series
I am building a model to forecast some metrics. Those metrics are quite seasonal giving me good forecasts as shown below:
However, some new requirements dictate that I target those forecasts per ...
0
votes
0
answers
34
views
Resources for Probabilstic forecasting
I was planning on learning probabilistic forecasting and im completely lost. Suggest some online resources. I have started with Probabilistic Forecasting and Bayesian Data Assimilation but its a bit ...
1
vote
1
answer
42
views
Is there anything inherently wrong to use the Black-Scholes-Merton model to simulate BTC pricing rather than only using it for Options pricing?
My understanding of the Black Scholes model is that it can be used to simulate options pricing for stocks in traditional financial markets. But this article uses the model to forecast/simulate bitcoin ...
1
vote
0
answers
33
views
Prediction based on new data in the tsDyn package [closed]
I decided to use the tsDyn package to build the model and precision the values. The argument for this package (I previously worked with urca, vars) was the ability to make predictions based on new ...
0
votes
0
answers
23
views
Prediction when Target's lag values are part of Predictors
I'm using LGBM for regression, where the Target column's lagged values (7 columns for each lag day) are also used as predictors when training the model. Absence of the 7Day lag values severely ...
0
votes
1
answer
88
views
Preventing Data Leakage in Time Series Forecasting with Feature Engineering
In a previous question (linked here), I sought guidance on forecasting thousands of time series. Based on the suggestion to treat it as a regression problem, I used the LightGBM model with extensive ...
0
votes
0
answers
9
views
How to handle recent variable change resulting in level-switch in time series modeling?
I developed a script to run time series models on people data, and I re-run the Arima model fitting/forecast reconciliation algorithm monthly as new data comes in. I use the grouped/hierarchical time ...
1
vote
1
answer
35
views
Bad performance of ARIMA model on online buzz data, Any suggestions?
I was wondering if the ARIMA model is constrained to predict online buzz data (time series data).
What I want to do: Use the past round 30 months data to predict next month; and I use Python
Here are ...
0
votes
0
answers
4
views
Should I sample or compute the innovation term in the recursive estimation OLS algorithm applied to time-series data?
I have time-series data $\{Z_{[t]}\} = Z_t, Z_{t-1}, ..., Z_1$ of length $L$, and I would like to estimate a model for the $\{Z_{[t]}\}$ so I can forecast $Z_{t+1}$ as a function of a number of its ...
0
votes
0
answers
25
views
auto.arima (Hyndman R package)
I am running an auto arima on a datase that yields two tries as revealed by using trace=TRUE as:
...
0
votes
0
answers
18
views
How to construct forecast confidence interval from the historical RMSE?
I'm reading a FED paper on forecast uncertainty estimation. Here's the link
One of their methods to estimate forecast interval is illustrated in the following figure
According to the description: &...
0
votes
1
answer
48
views
Can I use the residuals of a time series decomposition to estimate the effect of a covariate?
Context
I work for a company that has an e-commerce website. Regularly we make specific campaigns in order to sell more. For example: We can make a campaign for fathers day, black Friday, crazy August,...
0
votes
0
answers
42
views
Multiple comparison problems for Unit Root testing
Lets say I have a multivariate time series and I will run unit root tests for each variable to figure out if the variable is stationary and how many differencing are necessary to make it stationary.
...