Questions tagged [arima]

Refers to the AutoRegressive Integrated Moving Average model used in time series modeling both for data description and for forecasting. This model generalizes the ARMA model by including a term for differencing, which is useful for removing trends and handling some types of non-stationarity.

Filter by
Sorted by
Tagged with
0
votes
1answer
13 views

Motivation behind GARCH

Suppose I have built an ARIMA model for a real life process, where volatility is present. Now if for modelling the ARCH effect I fit a GARCH model, how will it affect my ARIMA model in terms of ...
0
votes
0answers
10 views

How to find the right parameters for the SARIMA model?

Here is the ACF and PACF graphs for differentiated series: From the ACF graph it looks like AR() model is more suitable, because the lags are decreasing exponentially and they show that the series ...
-1
votes
0answers
12 views

I am having difficulty in interpreting ACF and PACF plots ? (In ARIMA Model (p,d,q) ? I am doing emissions forecasting for annual data [closed]

ACF and PACF Plots are generated using differenced and natural log applied time series. Code is : acf2(dlnco2emission.ts, max.lag = 40) - applied in R Observations are 50 (yearly) values from EDGAR. I ...
0
votes
0answers
7 views

Multiple models walk forward validation ARIMA

I am trying to build a robust model to predict attrition rates based on 88 data points.I came across walk forward validation. I have the following understanding of it: We can either use the same set ...
3
votes
1answer
92 views

Python: auto_arima predicts constant value

As a newbie, I am trying to implement the forecast using the auto Arima model. After searching, I found this site illustrates the usage and the hyperparameters used in the model. However, when I tried ...
1
vote
0answers
61 views

Can vanilla NN model a moving Average (MA) process?

Let say I simulate an AR(1) process. We can easily model the data with a vanilla NN with a single neuron as it is about finding the linear relationship between $y$ and $y_{t-1}$. Now, how about a MA(1)...
0
votes
0answers
20 views

'low limit of 30% confidence interval' is higher than 'forecasts' in arimax model

I ran the ARIMAX model for multi-step forecasts (steps=n). The code was written in python and I used the ARIMA model from statsmodels library. (https://www.statsmodels.org/stable/generated/statsmodels....
0
votes
0answers
25 views

Forecast is not deterministic for statsmodels statsmodels.tsa.statespace.sarimax.SARIMAX

I'm forecasting a data time series, but each time I do it, I get a different forecast, sometimes with variations of more than 5%, which is not considered acceptable. The model Im using is v0.12.2 ...
0
votes
0answers
19 views

Identifying $ARIMA(p,d,q)\times(P,D,Q)s$ process

I have a series of monthly data (top left) differenced (12) with lag = 1 (bottom left) with the following ACF (top right) and PACF (bottom right): And I'm trying to fit an $ARIMA(p,d,q)\times(P,D,Q)s$...
0
votes
1answer
28 views

Standardized residuals check in a ARMA-GARCH model

If standardized residuals of an ARMA-GARCH model show some autocorrelation, while the squared standardized residuals look white noise, what can we infer about the specification of the model?
0
votes
0answers
10 views

How to determine the parameters for SARIMA(P,D,Q,p,d,q) model using ACF and PACF graphs?

I'd like to apply SARIMA model to my time series. I'd like to know how I could determine/guess the parameters by using ACF and PACF graphs. The time series data range from years 1995 - 2020. The data ...
1
vote
0answers
40 views

Is ARIMA and Random Walk a Nested model?

I have confused about the Radom Walk. If a model restrict some parameter as 0 and the two regressions are the same, which is the Nested model. But in ARIMA model versus Random Walk model, the random ...
0
votes
0answers
18 views

How to determine the lag.max length for daily (10 years) data stock prices data set in R

I have 20 years of daily data set for some stock prices. I would like to investigate the autocorrelation for this using acf plot and ...
0
votes
0answers
37 views

Time Series Forecasting Arima family methods

I'm fairly new to time series analysis and forecasting. I'm using the uci househould power consumption dataset to build a model to forecast energy consumption. The dataset measures the power (kW) ...
0
votes
0answers
12 views

Is SARIMAX dependent on the scale of an exogenous variable?

I built a SARIMAX model that forecasts some sales. The model includes one exogenous variable (the number of people on holiday). I used this implementation: https://www.statsmodels.org/stable/generated/...
1
vote
1answer
35 views

Using R auto.arima and arima.sim for stock prices

I simulate around 16000 stock prices with using auto.arima and arima.sim and I have two questions. Do I need to use plain ...
1
vote
0answers
16 views

Exogenous variables SARIMAX

I am trying to model some data as a time series analysis, and I have one exogenous variable that I have reason to believe should have a significant effect on the response. but when I include the ...
2
votes
1answer
34 views

Getting flat prediction with ARIMA model in Python

I am using an arima model to forecast sales of a given product in python, using statsmodels.tsa.arima.model.ARIMA Sales are daily, with a history of 2019 until ...
1
vote
0answers
22 views

Deriving expression for Variance in Time Series (example)

Given an $ARIMA(1, 2, 0)$ process $X_t$, let $\nabla^2X_t = Y_t$, then $Y_t = \alpha Y_{t-1} + \epsilon_t$ is a stationary $AR(1)$ process. (Here $\nabla$ is the difference operator such that $\nabla ...
0
votes
0answers
12 views

Looping with arima in R [migrated]

I am trying to do multiple arimas with the for function. So far my attempt is this. ...
0
votes
0answers
37 views

Time series model to analyze hard disk failure

I wanted to use time series models to analyze the Backblaze Hard Drive Data. Here SMART features are associated with the hard drive failure. In the failure column, 0 represents healthy drive, and 1 ...
1
vote
2answers
62 views

Express an I(2) process as a sum of telescoping cumulative sums

Suppose $z_t$ is zero mean I(2) process Show that $z_t$ can be written as $$z_t = \sum_{i=1}^t e_i +\sum_{i=1}^{t-1}e_i +\sum_{i=1}^{t-2}e_i +...+e_1 $$ where $e_i$ is white noise process.
1
vote
0answers
37 views

Forecasting: AIC, AICc and BIC VS Cross Validation for model selection (cases for different horizons)

The majority of the automatic model selection algorithms like auto.arima and ets (https://robjhyndman.com/publications/automatic-...
0
votes
0answers
8 views

Investigating lagged effects of an independent variable on a replicated dependent variable

I have data from repeated observations of ~60 independent replicates over 20 consecutive years, along with an environmental index. Let's say these data are the annual quantity of apples produced by 60 ...
0
votes
0answers
19 views

ARIMA Correlogram Classification - Insignificant P-Values (Reposted w/ specific questions) [duplicate]

I am trying to apply ARIMA forecasting to a stock's market close price. It is daily data with 129 observations. I used Augmented Dickey-Fuller Test and a Correlogram to confirm the data is non-...
0
votes
1answer
33 views

Parameter estimation of ARIMA model with exogenous variables (ARIMAX)

I am trying to compare an ARIMA model based on the price of a cryptocurrency without exogenous variables to one which adds in the number of tweets about the crypto in the same period as an exogenous ...
0
votes
1answer
34 views

ARMA and AR processes

I am taking time-series econometrics this semester and got stuck with the following. Assuming we have $ARMA(1,1)$ model: $Y_t = 0.2Y_{t-1} + ε_t + 0.1ε_{t-1}$ with the estimated variance of $1$. ...
7
votes
2answers
467 views

Why is AIC or BIC commonly used in model selections for time series forecasting?

On scikit-learn documentation, I found the following comments about AIC: Information-criterion based model selection is very fast, but it relies on a proper estimation of degrees of freedom, are ...
0
votes
1answer
17 views

Parameter simplification of ARIMA model

I am constructing an ARIMA model on a cryptocurrency price time series. Using the autocorrelation and partial autocorrelation plots I came to the parameters of (p,d,q)=(3,1,2). The resulting RMSE was ...
0
votes
1answer
17 views

How can I account for large lag cycles in timeseries regression with ARIMA errors?

I'm trying to create a model and generate a forecast of energy consumption in an HVAC system at the university I attend. I have energy consumption for the system and some basic weather data. Sample of ...
0
votes
0answers
14 views

Find the order of integration of an ARIMA model from a specified equation

Suppose I'm given an equation: $$ Y_t = 1.7 Y_{t-1} - 0.7 Y_{t-2} + e_t $$ If I'm given the data, I can use the Box-Jenkins methodology coupled with stationarity tests like the ADF test to find the ...
0
votes
0answers
14 views

Heteroskedasticity Arima Residuals with Xregs using Breusch Pagan

I know there are several related questions, but my question is very specific. If I want to test the heteroskedasticity of the residuals of an Arima model with external regressors, what formula should ...
0
votes
0answers
13 views

How to select ARIMA model with cyclic ACF?

My annual time series has following ACF/PACF structure. Based on what ARIMA model should be selected here? Exponential decreasing of ACF --> AR(4) probably? Or because of periodical ACF maybe SMA? ...
0
votes
0answers
28 views

SARIMA vs SARMA model

I've built two time series models using the same data. One model is a SARIMA (1,1,1) (0,0,0,52) model built with at level data and the second model is a SARMA (1,0,1) (0,0,0,52) model built with first ...
0
votes
0answers
19 views

Effect of scaling data on ARMA coefficients [duplicate]

For numerical stability, I thought it might be a good idea to scale my data before feeding them into an ARMA GARCH model. I have gone through a few older posts and understand the affect scaling ...
0
votes
0answers
13 views

Types of volatility predictors

I am slightly confused by this paper: https://www.researchgate.net/publication/228290046_Practical_Issues_in_Forecasting_Volatility In the paper they consider roughly four groups of models: historical ...
0
votes
0answers
24 views

ARMA Models with Trend Question

I am looking to build an ARMA model for a time series with a significant trend component. Let's assume for the purposes of this question that I don't want to build an ARIMA model. (The reason has to ...
1
vote
0answers
39 views

ARMA GARCH fitting

I've made a few posts regarding a manual ARMA GARCH implementation and I have made some great progress. However, I am still shy of a working program as I am obtaining some rather large forecasts. I've ...
0
votes
0answers
22 views

Orders of AR and MA models

I have a couple of ACFs and PACFs and cant seem to be sure about the order of AR(p) and MA(q).Can anybody kindly give me an insight into how to detect that?
1
vote
0answers
25 views

How can i use an ARIMA model to explain the an effect of news regarding COVID-19 on a stock market index?

My thoughts were to model the time series of the stock index up until the particular day for the news I am looking at and forecast using my model. Then I would compare the forecasted result to the ...
1
vote
1answer
33 views

How to find the order for ARMA model?

I have a problem finding the order with the ACF and PACF plot, below is it. First I think they can be considered as tails off gradually because they are abnormal, then I set AR(1) from PACF and MA(1) ...
1
vote
0answers
9 views

Are consecutive zeros across multiple dimensions in multivariate time series a problem when estimating VARMA models?

I want to estimate a VARMA model for a 14-dimensional multivariate time series (Fig. 1, 2). The goal is to investigate how the trajectory of my alleged output time series (messages per hour; Fig. 1, ...
1
vote
0answers
16 views

Why ARIMA Ljung-Box p-values tend to decrease for large lag

I am a beginner at ARIMA model fitting. I have observed that after I fit my model, the Ljung-Box statistic tend to discard the null hypothesis for large lag. It could be that it is only random that it ...
0
votes
1answer
40 views

Issues Manually Implementing ARMA GARCH

I have been working on manually implementing an ARMA GARCH (1,1) model but have been running into a few issues, namely a very large forecasted variance. I am hoping by outlining my process someone can ...
0
votes
0answers
248 views

MA(1) process by-hand step-by-step instructions : parameters, fitted values, residuals

I would be very grateful if someone would provide STEP-BY-STEP instructions on how to calculate the parameters mu & theta, fitted values and residuals of an MA(1) process by-hand. For example, ...
2
votes
2answers
174 views

ARIMA model with least AIC giving negative forecasts even though there are no negative values in the training data

I'm training a arima model on a daily time series data where I'm trying to forecast daily inflow counts of a request on a particular day. It can be either positive or zero (no zero requests in the ...
0
votes
0answers
16 views

how does arima calculate phi, theta and some more

this is my first question i have ever asked on here, so sorry if its asked somewhere else. I made a random test dataset containing values: 10,4,8,5,6,4,7,8, and i tried an SARIMA(1,0,1)(1,0,1) and i ...
0
votes
1answer
25 views

Where do the error terms come from in ARMA?

I understand ARMA is a linear combination of lagged data points and lagged errors, but I am unclear on its implementation once parameters have been identified. Now suppose I have an ARMA model and ...
0
votes
0answers
28 views

Fitting ARMA GARCH

I am interested in fitting an ARMA GARCH model by hand (that is without the use of a package such as rugarch), but am unclear on how the parameters are estimated. I have read that one should use MLE, ...
0
votes
0answers
14 views

Create stationary dataset with log, sqrt and cube root layering

I have a time variant dataset with highly non stationary data. The dickey-fuller p value always at the 0.90 to 0.99 area. Even after single transformation, still data generates with p value about 0.80....

1
2 3 4 5
56