Skip to main content

Questions tagged [differencing]

Differencing is a time series transformation used for removing unit roots. It can be simple or seasonal (for seasonal unit roots), first-order or higher-order (for multiple unit roots), also fractional order. Do NOT confuse with tag *differences*

Filter by
Sorted by
Tagged with
3 votes
1 answer
50 views

Estimating an ARMAX model using an I(2) exogenous variable

Is it valid to estimate an ARMAX model using I(1) and I(2) variables, which are made stationary after first and second differencing, respectively? For instance, I have an I(1) stock price variable, ...
Pepe Frog's user avatar
0 votes
0 answers
54 views

Is my approach to compute Granger causality valid?

I have two time-series, let us call them A (colored in red) and B (colored in blue). There are ~770 data points per time-series. Note: Both time-series are in fact not the recorded raw signals, but ...
Philipp's user avatar
  • 307
2 votes
1 answer
159 views

Non-stationary time series: what are the advantages of doing analysis in levels instead of differences?

Suppose we want to analyze some non-stationary time series, x(t) and y(t). For simplicity, assume they are I(1). We can analyze them in levels (using cointegration tests) or in differences. What are ...
James's user avatar
  • 2,742
0 votes
1 answer
26 views

Why the differenced at lag 12 time series of a SARIMA(0,0,0)(0,1,1)_12 model follow the MA(1) pattern with step 12?

I am trying to understand why the ACF of the seasonally differenced series reveals the AR of MA structure of the original series. For example: The following lines creates a SARIMA(0,0,0)(0,1,1)_12 ...
Epameinondas's user avatar
0 votes
1 answer
54 views

Only first differencing the independent variable?

I am trying to study the relationship between income and preferences for redistribution. I have a panel dataset. I can simply the model (1): preferences = alpha + income + e. However, I'm also ...
chunguc1004's user avatar
5 votes
1 answer
71 views

Why does differencing White Noise induce autocorrelation of $-0.5$?

I am curious about the following problem. Let's have a variable given by white noise, $$y_t \sim \operatorname{NID}(0,1).$$ Let's say we difference it, $$\Delta y_t = y_t - y_{t-1}.$$ And now, if we ...
Athaeneus's user avatar
  • 205
0 votes
0 answers
19 views

Different results when fitting ARIMA model for levels vs ARMA model for first differences in R

In the following code I show that I get different forecasts when fitting an ARIMA(2,1,0) for cumulative sums of a generated AR(2) model vs. fitting an ARMA(2,0) for the AR(2) itself. Can anyone point ...
Mr Frog's user avatar
  • 339
0 votes
0 answers
6 views

Difference transformation and Stationarization of Moving Average

I have a temperature sensor data, I want to denoise it. The first thing that came to my mind was to take the moving average, it was very smooth but it is still not stationary. If I take the log ...
Clankk's user avatar
  • 33
1 vote
1 answer
20 views

Estimating VAR of differences of potentially cointegrated variables

What are the possible issues, if any, of estimating a standard two-dimensional $VAR(p)$ of $I(0)$ variables that are first differences of $I(1)$ variables whose potential cointegrating relationship ...
Pavel Filip's user avatar
2 votes
1 answer
81 views

Intuition behind the first, second, third order differencing. When to use them?

I am working with time series data and out of curiosity took the first, second and third order differences of my data. However, I am not quite sure when they are used? I also wonder about the ...
Mostafa Bouzari's user avatar
0 votes
0 answers
45 views

How to choose the order of differencing?

I have a time series that looks like this I was asked to display the autocorrelation function and perform ADF test and based on that suggest a d for further analysis. I get plots like this and can ...
moon's user avatar
  • 1
1 vote
1 answer
42 views

Auto arima capital D

I've got time series 3 years long, there is seasonal uplift during December - but it's not so clear. Seasonal test fails. I train model twice without setting any parameter: ...
voncuver's user avatar
5 votes
1 answer
183 views

Role of `trend` argument compared to integral order in ARIMA model

I am currently studying ARIMA models. When I checked for a Python library to train one, I stumbled upon statsmodels which features ARIMA (and SARIMAX from which ...
Marco Bresson's user avatar
1 vote
0 answers
61 views

Specifying parameters for SARIMAX model with significant ACF / PACF at tails

I have hourly data that has a period of 1 day or 24 hours / time steps and I hope to do short term forecasting for a few days in advance. The ACF of the raw time series was periodic (see last figure) ...
Yandle's user avatar
  • 1,189
0 votes
0 answers
41 views

How to interpret those ACF & PACF?

I have some problems when analyzing my time series dataset. Basically, the dataset is about the daily sales volume of an FMCG company (they work from Monday to Saturday with Sunday being a day off, so ...
Khoi Le's user avatar
0 votes
0 answers
21 views

ADF & KPSS Test Reporting Series as I(19)

I am currently attempting to determine the order of integration for a nonstationary time series of the Federal Funds Rate (https://fred.stlouisfed.org/series/FEDFUNDS - Monthly, 1990 to 2004, 168 obs.)...
shrey_shankar's user avatar
0 votes
0 answers
15 views

Differencing and Stationarity in Panel Data Analysis

As far as I know, the source of randomness is different in cross-sectional and time-series analysis. The following is my understanding. When we use a cross-sectional data set $\{X_i\}_{i=1}^{N}$, the ...
MinChul Park's user avatar
2 votes
0 answers
87 views

Applying differencing for time series data before regression analysis

I am new to time series regression. I am trying to understand the effect of differencing the time-dependent data before applying regression analysis. I tried fitting lots of regression models but in ...
Nipuni Opatha's user avatar
0 votes
0 answers
35 views

Gradient vs differences to remove non-stationarity in time series?

When dealing with non-stationary time series (for instance, in auto-correlation analysis), differencing (computing absolute differences between consecutive samples/observations) is often regarded as ...
joaocandre's user avatar
1 vote
0 answers
33 views

SARIMA model selection for my data

I am new to time series analysis. I need to determine whether my series is seasonal or not, and if it requires differencing for building an ARIMA model if it is possible? The time series data is ...
Antonio's user avatar
  • 11
3 votes
1 answer
58 views

If $X_t$ is an AR(2) process, what is $Y_t := X_t - X_{t-1}$?

Q: If $X_t$ is an AR(2) process, what is $Y_t := X_t - X_{t-1}$? Attempted solution: $X_t = \phi_1 X_{t-1} + \phi_2 X_{t-2} + W_t$, where $W_t$ is white noise. \begin{equation} \begin{split} Y_t &:...
Oskar's user avatar
  • 265
0 votes
0 answers
16 views

What happens if you difference a IMA?

When we over-difference a time serie, we are introducing units roots in its moving average component, hence obtaining an IMA. My question is: IMA is an integrated time serie, hence it has unit root ...
Giacomo Gregori's user avatar
0 votes
1 answer
34 views

Can I transform variables to comply with conditions on order of integration in an ARDL model?

If log variable have a unit root, can you then difference it? For ARDL model variables should not be I(2) or more. Variables of log form are I(2). Would it be a problem to difference it. What I am ...
Gus's user avatar
  • 25
1 vote
2 answers
75 views

First difference of logs of negative numbers causes trouble

I have the following problem: I have a timeseries with the prices for a few futures, which is non-stationary (according to ADF test). If I apply first difference of logs, ADF shows stationarity. But I ...
Arri's user avatar
  • 47
1 vote
1 answer
38 views

Direct multi-step forecasting an integrated time series from a stationary model

I am new to time series analysis. But I am well versed in data science and machine learning. One thing that confused me was modeling the data. For example, I want to create an LSTM or regression model,...
Clankk's user avatar
  • 33
0 votes
0 answers
47 views

ARDL Unit root df

When applying ARDL I used Dickey-Fuller unit root test, and found variables integrated at I(1) and I(0). For running the ARDL model do I use my original data (not differenced) or do I use the ...
Gus's user avatar
  • 25
0 votes
0 answers
314 views

Interpretation of DCC-GARCH model

For my Master thesis I have to perform the DCC-GARCH model to examine the correlation between real estate house prices and the stock market. I tested the data for normality (both not normal) and ...
Jan-Willem van Boven's user avatar
3 votes
2 answers
572 views

What's the 'right' number of parameters for an ARIMA model?

I'm working on an unassessed course problem, The file Pas-mile.txt contains the monthly numbers of passenger miles travelled on US airlines for each month between ...
mjc's user avatar
  • 599
0 votes
1 answer
60 views

Modelling temperature: seasonality, stationarity, differencing

I have temperature-related time daily time series. I plot the time series and found that the plot have seasonal variations. Thus I differenced the series. When I did Dickey-Fuller test for both ...
soba's user avatar
  • 3
1 vote
0 answers
35 views

How to generate quantile forecasts from first differences?

Let's say I have a time series and I am taking the first differences and training a model to output the predicted 95% quantiles of these first differences at future time horizons. If this was just a ...
iYOA's user avatar
  • 111
1 vote
1 answer
44 views

ARIMA D coefficient for seasonality lesser than frequency

Lets suppose we have a time series with monthly data (frequency=12) my_tS <- ts(my_monthly_data, start=c(2002,1), frequency=12) When I plot it, I can see a ...
Kaikus's user avatar
  • 219
3 votes
0 answers
53 views

Differencing Long Memory Time Series

I came accross some articles stating that differencing a long memory time series leads to memory loss. I don't know much about long-memory time series, but I know how $ARMA$ and $ARIMA$ processes work....
Residual Claimant 's user avatar
0 votes
0 answers
233 views

How to choose model parameters based on ACF and PACF in SARIMA model?

I have some monthly data which show the number of visitors to a country whose plot is given below Since there seems a non-constant variance, I took the log, which stabilized the variance. To remove ...
Günal's user avatar
  • 951
0 votes
0 answers
55 views

Is the first difference of a strongly stationary process stationary?

Let $X_t$ be a strongly stationary time series. Is the first-order difference process $\nabla X_t$ always stationary?
Dime's user avatar
  • 125
0 votes
1 answer
110 views

Linear regression with ARIMA errors and seasonal dummy covariates: how does differencing works?

To model my daily time-series data, I want to use linear regression with ARIMA errors. I also want to introduce several seasonal dummy covariates (day of the week, month of the year). I read in ...
adrimsvieira's user avatar
0 votes
0 answers
45 views

AR Modelling. Box Cox and Differencing don't give Stationary Data

I am trying to fit an AR/ARIMA model on electricity data on hourly prices during a almost three year period, I am following the guidelines in Hyndman, R.J., & Athanasopoulos, G. (2018) to do this. ...
Simon Rydstedt's user avatar
1 vote
0 answers
126 views

How to calculate the forecast interval for time series prediction obtained by doing seasonal differencing before fitting arima

Here is the link of my previous question. How to forecast a time series which is generated by accumulating data of every five minutes and reset to 0 by the end of the day I am working with a seasonal ...
Charles Zhou's user avatar
0 votes
1 answer
456 views

Why Python does not seem to correctly report fitted values in statmodels ARIMA when differencing is involved?

When fitting an ARIMA model using the statsmodels python implementation I see the following behaviour, python does not seem to correctly provide the values for the differenced lags. I am comparing the ...
Alberto GR's user avatar
0 votes
0 answers
30 views

Fit an autoregressive model on a heteroscedastic time-series

I have a time-series that I want to model using an autoregressive process using statsmodel for benchmarking purposes. This time-series is not stationary: visual ...
Dime's user avatar
  • 125
1 vote
1 answer
33 views

Sum of $I(1)$ and WN

Is the process $\Delta Y$ an MA(1)? $Y_{t} = X_{t} + w_{t}$ with $X_{t} = X_{t-1} + e_{t}$ and $e_t$, $w_t$ both independent white noise a MA(1) process? What I did is the following: $(1-L)X_{t} = e_{...
JPbet's user avatar
  • 11
1 vote
0 answers
22 views

Is non-stationarity an issue in this setting?

I am trying to model the development of European spot prices of gas. My purpose is to explain what has caused past movements in the gas price, rather than forecasting future gas prices. Naturally, ...
Stefan's user avatar
  • 11
0 votes
1 answer
75 views

Reverting differencing for Time Series VAR model component wise?

I have data with seasonality so I used seasonal and then first order differencing to make the series stationary in order to fit VAR model. After the model is fitted on the differenced level I can ...
Lukas's user avatar
  • 3
1 vote
1 answer
52 views

For time series forecasting using machine learning, is feature engineering only needed for the y variable or is it needed for all x variables also?

Say I am trying to predict house prices (y variable) using population growth and GDP (x variables) using XGBoost or Neural Networks. All 3 are time series. I understand that I have to feature engineer ...
oldradishsoip's user avatar
5 votes
1 answer
2k views

For time series, is non stationary an issue for machine learning models like they are for traditional time series models like ARIMA?

Sorry if the question seems basic. I understand that non stationary data is a big issue for traditional time series forecasting methods like ARIMA and VAR but is it the same for machine learning ...
oldradishsoip's user avatar
2 votes
0 answers
74 views

Percentage change and then Rolling average V.S. rolling average and then percentage change

For a timeseries data $\{D_t\}$ with some seasonality $s$ and high volatility $\sigma$, we often want to perform a rolling average $\frac{1}{n}\sum_i D_{t-i}$ to denoise the data and take some ...
The One's user avatar
  • 225
0 votes
0 answers
110 views

How $\psi_{j}$ weights of the AR(p) process satisfy the difference equation?

The parameters $\phi_{1}, \phi_{2}, \ldots, \phi_{p}$ of an AR(p) process $$ z_{t} = \phi_{1} z_{t-1} + \cdots + \phi_{p} z_{t-p} + a_{t} $$ or $$ (1-\phi_{1}B-\cdots-B^{p}) z_{t} = \phi(B) z_{t} = ...
J.H's user avatar
  • 23
0 votes
0 answers
136 views

Show that if $x_t$ is a stationary time series then $\Delta x_t$ is also (weakly) stationary

I need to express $E[\Delta x_{t}]$ and $\gamma_{\Delta X}(h)$ in terms of $E[x_{t}]$ and $\gamma_{X}(h)$. The first part of this is trivial and can easily be shown that $E[\Delta x_{t}] = \mu_{X} - \...
rst231's user avatar
  • 1
0 votes
0 answers
134 views

How do I turn the n-ahead ARIMA predictions to undifferenced values?

I`m doing a Monthly sales forecast based on some historical data. My question is when I do the Arima(p,d,q) and then I do forecast(arima_model) in order to get the fitted values, which are nicely ...
Ivan Ar's user avatar
0 votes
1 answer
767 views

How to invert differencing in time series data if I am making multiple steps prediction?

I have a time-series that I would like to use for predicting 36 timesteps in advance using LSTM. It is not stationary so I differenced the series by subtracting each point from the next one. My ...
Traveling Salesman's user avatar
2 votes
1 answer
880 views

Should I difference my data before run a ADF test?

I plot my data as shown in the following screenshot: Clearly the series contains a trend. A first order difference of my data will eliminate the trend, which I plot as follows: Now I would like to ...
zyy's user avatar
  • 125