Hello Everyone! Thanks for Watching! 🌐Video 1: ARIMA models in STATA - Part 1: Identification Link: ua-cam.com/video/pquD3OjeLFU/v-deo.html 🌐Video 2: ARIMA models in STATA - Part 2: Estimation Link: ua-cam.com/video/mPDNH-rA4OQ/v-deo.html ✅ Patreon: Support my channel for more content creation and get special deals and members bonus (such as help in your research and material downloads): www.patreon.com/JDEconomics ✅ Get the DO files + Slides + Dataset at: payhip.com/b/Mj2v ✅ The tutorial is also available in EViews: ua-cam.com/video/ukGJ0sLgbqI/v-deo.html ✅ Subscribe to my channel by clicking: ua-cam.com/channels/5P21WGFO4WRUlAiGLcwymg.html ✅ You can get access to all the EViews Workfiles, DO files (STATA) and Slides from my videos at: payhip.com/JDEconomics Thanks a lot! JD Economics.
Hello. Can we do this forecasting for the ARIMAX model which refers to a set of (exogenous) independent variables for a single regression (and single dependent variable)?
I have watched almost all your videos and they are good. I used your two videos on VAR model using Stata and I literally followed your guide for my time series project. I am now studying for my exams and your videos have made this course more fascinating than my course at uni has done. Thank you sir!🙇♀️
Thank you so much for your videos on ARIMA. I learned so much and now can use it to write my paper. Many thanks and please keep on making more videos. Will be waiting for them.
Hello Zerin! Thanks for your message. I am happy to hear my video was able to help you! Definitely feel free to subscribe to get notified of new tutorials and feel free to share my UA-cam Channel with anyone you think may be interested! Best regards and good luck on your paper! JD
My pleasure! Thanks for your positive feedback! Please feel free to share my channel with your close ones and don’t forget to check my website. Www.jdeconomics.com Best regards, JD
@@JDEconomics You are doing a great job. Please, I have a question. If a time series variable has unit root, and the autocorrelation and the partial autocorrelation graph of the 1st difference of the render all lags within the confidence bound, how would the orders of the AR and MA component of an ARIMA model for such a variable be determined? Would it be necessary to consider the 2nd difference variable?
Thanks for your message and feedback! I will be submitting more videos shortly! Next topic is cointegration. Feel free to subscribe for more content! Regards, JDEcon.
Thank you for the detailed explanation of videos. I have followed all the method as suggested in your three videos. I have used the command "predict fbn, y dynamic(2022)" as I need to predict the yearly data of my variable from 2022. It did generate a new column "fbn" but it did not generate any values from 2022 to 2031 for ten years and all the rows were blank. I am not sure what was the issue. Could you please suggest me with any solution to fix this issue. Thank you so much.
Hello, greeting from Indonesia. Thank you so much for your videos on ARIMA. It helped me so much, however when I run same command ( in forecast part ), I didnt get the same result like yours. a forecast line, isn't behave like actual, and it shows resid only. could you please advise ?
The command may be running on the differenced series. I suggest you estimate the model, use the predict command after the arima command, and use "predict y,y ". See if that works. However, I do suggest you to check the arima posestimation manual which has a lot of useful information: www.stata.com/manuals/tsarimapostestimation.pdf Best Regards, JDEcon.
No worries! there is no easy way to generate the confidence bands. You can always get the standard error of the forecasts and create the upper bound and lower bound manually. Yet, it doesn't work great. Regards,
Hi thank you so much for the series of ARIMA videos you have shared. They help me a lot on my dissertation. But one thing to ask, after we finally find the best ARIMA model from stage 1-3, so where is the continuity from having a good ARIMA model to the forecasting process? Am I missing something here? Thank you so much!
Hi, thanks for your feedback. There are only 3 tutorials. The arima that has the most desirable properties among all of those you identified is the one you should use to forecast. Regards, JD
Hello JDeconomics. What if I have the import of beef as a dependent variable and my independent variables are the production of beef and the exchange rate. I am using time series annual data. Should I take into account all of them to forecast or only I can focus on my dependent variable only?
Dear Juan. If I want to forecast a dependent variable by adding factors (independent variables)that can affect the dependent variable, which command should I use to predict? Are the identification and estimation processes the same as not including independent variables? Thank you so much for your help. 🙏🙏🙏
Excellent video ... Could you make other videos about ARCH, GARCH, EGARCH, etc, and the Vector Autorregressive Models??? Thanks a lot about this very useful theme.
Hello Carlos, Thanks for your message. I have a video about Vector Autorregressive models in Stata. The link is: ua-cam.com/video/I2IpCRfi7ts/v-deo.html The tutorial is also available in Eviews. I will get into the other topics shortly (Garch, Arch, etc). Lastly, feel free to check my website: www.jdeconomics.com where you have all the tutorials available in an organized way so you can find them easier. Subscribe to my channel (if you haven't) so you get notified of the new videos. Kind Regards, JD.
Hello. Thank you so much for your videos! they're helping me with my dissertation. One question please; when we perform forecasting do we use the raw data of cpi or the first difference of it? i.e. while typing the command arima cpi, arima(1,1,1). is cpi the normal one or d.cpi? thanks a bunch.
Hello! Thanks for your comment. It’s doing it on the raw data. You can use the command xb (instead of y) to obtain the differenced mean equation. Here is the link for the predict command of ARIMA models. www.stata.com/manuals/tsarimapostestimation.pdf . Best regards, JD
@@JDEconomics Thank you for your timely reply! Last thing please. if our series is non-stationary, and we find that taking the first difference makes it stationary, we will have to run the arima with d.cpi right? or do we stick with cpi.
@@aleesh30 no, because if you use d.cpi and for example arima(1,1,1), you are telling stata to difference once the variable d.cpi. Meaning: You are differencing a differenced variable. I hope that helps to clarify)
Thanks for your feedback! I create the content of my tutorials. The literature I use to teach the content is displayed in my website. You can check the recommended literature at the end of the tutorial in the web. Best Regards, JD
Hi, thanks for your message. You mean the arima command or the predict command? You can just try using the following command, "predict y, y". That should create the forecasts. Also, feel free to check the arima postestimation manual which will give you some insights: www.stata.com/manuals/tsarimapostestimation.pdf Good luck! JDEcon.
Hi, Your videos arehighly informative and helpful. I have a small doubt, for the predict fcpi,y dynamic(m(2021m3)) i have is a unit. So how do i change the code and run? i tried - predict fnet_income(g(3069)), but the g() is not recognised. Can please tell me how to run the code if we have units instead of quarter or month time format.
Hello Karin, I haven't really worked in the past with climate change papers so I am not too familiar with the databases. I know the world bank database has CO2 emissions, which is normally used for climate change research. However, the frequency is yearly. Here is the link: data.worldbank.org/topic/19 Feel free to subscribe if you would like to get notified about the new content I submit! Regards, JD
The only thing I can think of are: 1- Make sure that estat aroots is a command that exists in stata 12. 2- Make sure that you are using the command right after estimating the arima nodel. Regards
Hello, Juan. Thanks a lot for your videos! Could I ask you please how could be written a command for ARIMA with raw data, if I made the logarithm to stabilize the variance ? Where "Lncpi" could be specified ?
Hi, I am not sure I understand your inquiry. But I think this is what you ar looking for: IF you use the command: *predict, fcpi y* , stata will forecast the variable in levels (regardless of the transformation you did on the data). However, if you use the command: *predict, fcpi xb* Stata will show the forecasts for the transformation. It is all in the manual, in page 3 you have further details. I attached the link for simplicity: www.stata.com/manuals13/tsarimapostestimation.pdf#tsarimapostestimation Regards, JD
@@JDEconomics Thank you so much, Juan. Your videos and explanation are really very very helpful, thank you very much! Best wishes for the upcoming holidays 🧑🎄
Thank you so much for your video. One question please: because I use the daily data and I set my time with 「gen t =_n」so that my new time series data would be 「t」 and that it starts from 1,2,...,900. How do I type the 「y dynamic(_(__))」about the underline in these two parts？ Thanks
Hi, Thanks for your message. There are only 3 videos, one for each stage of the Box Jenkins Method. What did you want me to cover in video 4? Regards, JDEc,
@@JDEconomics I greatly appreciate your immediate response. It would be great if there would be a tutorial on the forecast command and not the predict command. Or are they just practically the same? Because it seems like different ARIMA models would result to the same predict results. Thank you. Subscribed.
@@naokimikoshiba667 hi! Thanks for your message and subscribing! So you use the predict command to forecast periods ahead. Basically the predict command is used after a model estimation. By typing predict “y” we are telling Stata it has to predict the dependant variable (in our case cpi) and using dynamic we are telling Stata it needs to be a dynamic rather than static prediction. For this example, the graph will look very similar among diverse arima models specifications because as you can see the CPI is an increasing line. Any arima model for this case will fit a line. However if you type “browse” you will see that different arima models produce different predictions numbers. I hope that helped! Regards, JDEc.
Hello Everyone! Thanks for Watching!
🌐Video 1: ARIMA models in STATA - Part 1: Identification
Link: ua-cam.com/video/pquD3OjeLFU/v-deo.html
🌐Video 2: ARIMA models in STATA - Part 2: Estimation
Link: ua-cam.com/video/mPDNH-rA4OQ/v-deo.html
✅ Patreon: Support my channel for more content creation and get special deals and members bonus (such as help in your research and material downloads): www.patreon.com/JDEconomics
✅ Get the DO files + Slides + Dataset at: payhip.com/b/Mj2v
✅ The tutorial is also available in EViews: ua-cam.com/video/ukGJ0sLgbqI/v-deo.html
✅ Subscribe to my channel by clicking: ua-cam.com/channels/5P21WGFO4WRUlAiGLcwymg.html
✅ You can get access to all the EViews Workfiles, DO files (STATA) and Slides from my videos at: payhip.com/JDEconomics
Thanks a lot!
JD Economics.
Hello. Can we do this forecasting for the ARIMAX model which refers to a set of (exogenous) independent variables for a single regression (and single dependent variable)?
I have watched almost all your videos and they are good. I used your two videos on VAR model using Stata and I literally followed your guide for my time series project. I am now studying for my exams and your videos have made this course more fascinating than my course at uni has done. Thank you sir!🙇♀️
Great to hear! I wish you good luck at Uni! Cheers, JD
I am really thankful to you for uploading these videos on time series which helped me to get a new job as a research assistant. From India.
Thanks so much for your kind words! I wish you good luck! Best, JD
This video has been a life saver. Thank you so much!
Glad it helped!
Thank you so much for your videos on ARIMA. I learned so much and now can use it to write my paper. Many thanks and please keep on making more videos. Will be waiting for them.
Hello Zerin! Thanks for your message. I am happy to hear my video was able to help you! Definitely feel free to subscribe to get notified of new tutorials and feel free to share my UA-cam Channel with anyone you think may be interested! Best regards and good luck on your paper! JD
If I get an A in this course, sir, just know you were part of it. Watched all three videos. Thank you so much sir!!!
Great! Let me know how it goes! I wish you good luck. Please feel free to share my channel with your friends. Best, JD
Thanks a ton for these videos. All 3 of them were super helpful. You made my life easy. Very grateful!!!!
My pleasure! Thanks for your positive feedback! Please feel free to share my channel with your close ones and don’t forget to check my website. Www.jdeconomics.com
Best regards,
JD
I would like to thank you for this clear and helpful explanation for ARIMA process
My pleasure! Thanks a lot for your kind feedback! I wish you very good luck! JD
Thank you so much from Cambodia.
@@sralaophoneshop5028 no problem! Cheers
Impressive. Thanks for sharing the knowledge
Thanks for your kind feedback! Wish you good luck!
Thank you so much for such a brilliant video!! Easy to understand and follow. 100 thumbs up! 👍👍👍👍👍👍👍👍👍
Thanks a lot!
Many thanks for the videos. They are very helpful.
Thanks for your comment! Make sure to visit my website www.jdeconomics.com
@@JDEconomics You are doing a great job.
Please, I have a question.
If a time series variable has unit root, and the autocorrelation and the partial autocorrelation graph of the 1st difference of the render all lags within the confidence bound, how would the orders of the AR and MA component of an ARIMA model for such a variable be determined? Would it be necessary to consider the 2nd difference variable?
you are just great
Thanks!
Great videos, quite helpful and remarkably organised. It would be great to watch video on SARIMA model, how to identify (at least)
Thanks! And yes! Hopefully soon. Regards, JD
Excellent video! Thank you.
Glad you enjoyed it! Regards, JD
this video is amazing., thanks
Thanks for your comment! I am glad it helped you. Best regards, JD
Thank for your video, it's useful for me to finnish my study. I hope you can product more videos about economists such as Arch, Garch...
Thanks for your message and feedback! I will be submitting more videos shortly! Next topic is cointegration. Feel free to subscribe for more content!
Regards,
JDEcon.
Thank you for the detailed explanation of videos. I have followed all the method as suggested in your three videos. I have used the command "predict fbn, y dynamic(2022)" as I need to predict the yearly data of my variable from 2022. It did generate a new column "fbn" but it did not generate any values from 2022 to 2031 for ten years and all the rows were blank. I am not sure what was the issue.
Could you please suggest me with any solution to fix this issue. Thank you so much.
Commands estat aroots and
tsappend, add(5) are not working for me, what shall I do?
i also have same issue
very helpful tutorial video. Thanks a lot!
Great! Thanks
EXCELLENT
Thanks a lot! Please feel free to subscribe to my channel and share it with your friends/social network. I wish you good luck in your studies/work!
Thank you so much! very helpful video series!!
Thanks! Please feel free to check my website: www.jdeconomics.com
Regards!
Hello, greeting from Indonesia.
Thank you so much for your videos on ARIMA. It helped me so much, however when I run same command ( in forecast part ), I didnt get the same result like yours. a forecast line, isn't behave like actual, and it shows resid only. could you please advise ?
The command may be running on the differenced series. I suggest you estimate the model, use the predict command after the arima command, and use "predict y,y ". See if that works. However, I do suggest you to check the arima posestimation manual which has a lot of useful information: www.stata.com/manuals/tsarimapostestimation.pdf
Best Regards,
JDEcon.
Thank you! How can we add confidence intervals to the forecast?
No worries! there is no easy way to generate the confidence bands. You can always get the standard error of the forecasts and create the upper bound and lower bound manually. Yet, it doesn't work great. Regards,
Hi thank you so much for the series of ARIMA videos you have shared. They help me a lot on my dissertation. But one thing to ask, after we finally find the best ARIMA model from stage 1-3, so where is the continuity from having a good ARIMA model to the forecasting process? Am I missing something here? Thank you so much!
Hi, thanks for your feedback. There are only 3 tutorials. The arima that has the most desirable properties among all of those you identified is the one you should use to forecast. Regards, JD
@@JDEconomics okay i got it finally. So i need to generate the best model of ARIMA first before start to forecast. Thx Juan!
@@JDEconomics Hi Juan, do you have any example on doing Interrupted Time Series using ARIMA with multiple independent variables?
Hey, no. Sorry
Hello JDeconomics. What if I have the import of beef as a dependent variable and my independent variables are the production of beef and the exchange rate. I am using time series annual data. Should I take into account all of them to forecast or only I can focus on my dependent variable only?
Hi! An Arima model is not appropriate for what you want to do. Linear regression or var model. Regards
Dear Juan. If I want to forecast a dependent variable by adding factors (independent variables)that can affect the dependent variable, which command should I use to predict? Are the identification and estimation processes the same as not including independent variables? Thank you so much for your help. 🙏🙏🙏
rlly useful for me thx!
Great! Thanks
Excellent video ... Could you make other videos about ARCH, GARCH, EGARCH, etc, and the Vector Autorregressive Models??? Thanks a lot about this very useful theme.
Hello Carlos, Thanks for your message. I have a video about Vector Autorregressive models in Stata. The link is: ua-cam.com/video/I2IpCRfi7ts/v-deo.html
The tutorial is also available in Eviews. I will get into the other topics shortly (Garch, Arch, etc). Lastly, feel free to check my website: www.jdeconomics.com where you have all the tutorials available in an organized way so you can find them easier.
Subscribe to my channel (if you haven't) so you get notified of the new videos.
Kind Regards,
JD.
Thanks!! You are great!
Thanks Sona! I am glad you find it useful. Beat regards, JD
@JD Economics, how do I add 95% confidence interval to my forecasted model?
Hey! I don’t have the code, or know how to add a confidence interval to the forecast in Stata. Regards
Hello. Thank you so much for your videos! they're helping me with my dissertation. One question please; when we perform forecasting do we use the raw data of cpi or the first difference of it? i.e. while typing the command arima cpi, arima(1,1,1). is cpi the normal one or d.cpi? thanks a bunch.
Hello! Thanks for your comment. It’s doing it on the raw data. You can use the command xb (instead of y) to obtain the differenced mean equation. Here is the link for the predict command of ARIMA models. www.stata.com/manuals/tsarimapostestimation.pdf . Best regards, JD
@@JDEconomics Thank you for your timely reply! Last thing please. if our series is non-stationary, and we find that taking the first difference makes it stationary, we will have to run the arima with d.cpi right? or do we stick with cpi.
@@aleesh30 no, because if you use d.cpi and for example arima(1,1,1), you are telling stata to difference once the variable d.cpi. Meaning: You are differencing a differenced variable. I hope that helps to clarify)
@@JDEconomics omg yes! thank you. You just saved my thesis!
Lol. Glad to hear so. Feel free to get me a coffee if you do well! Lol. I wish you good luck! Best regards
How can we do BG test ARIMA after estimation?
I appreciate your tutorials. Please, I want to know the author of the textbook (Time series and applied econometrics).
Thanks for your feedback! I create the content of my tutorials. The literature I use to teach the content is displayed in my website. You can check the recommended literature at the end of the tutorial in the web. Best Regards, JD
Could you possibly help me with the error in stata (option yline() not allowed), u've done it in the part with white noise
Dear sir, how to write the command for daily forecast?
Hi, thanks for your message. You mean the arima command or the predict command?
You can just try using the following command, "predict y, y". That should create the forecasts. Also, feel free to check the arima postestimation manual which will give you some insights: www.stata.com/manuals/tsarimapostestimation.pdf
Good luck!
JDEcon.
Hi, Your videos arehighly informative and helpful. I have a small doubt, for the predict fcpi,y dynamic(m(2021m3)) i have is a unit. So how do i change the code and run? i tried - predict fnet_income(g(3069)), but the g() is not recognised. Can please tell me how to run the code if we have units instead of quarter or month time format.
Hey! Yes, you need to review the syntax. Looks like its wrong your code. Also, I don;t know what g stands for. Good luck!
@@JDEconomics g for generic likewise m for month and q for qurater.
it will be of great help, if you can help me.
can i help for to get data
topic thesis are
the impact of climate change on agriculural product thank you
Hello Karin, I haven't really worked in the past with climate change papers so I am not too familiar with the databases. I know the world bank database has CO2 emissions, which is normally used for climate change research. However, the frequency is yearly.
Here is the link: data.worldbank.org/topic/19
Feel free to subscribe if you would like to get notified about the new content I submit!
Regards,
JD
Hi, your videos are extremely helpful! just one thing, how do I add confidence band in the forecast in stata?
The only thing I can think of are:
1- Make sure that estat aroots is a command that exists in stata 12.
2- Make sure that you are using the command right after estimating the arima nodel.
Regards
Hello, Juan. Thanks a lot for your videos! Could I ask you please how could be written a command for ARIMA with raw data, if I made the logarithm to stabilize the variance ? Where "Lncpi" could be specified ?
Hi, I am not sure I understand your inquiry. But I think this is what you ar looking for: IF you use the command: *predict, fcpi y* , stata will forecast the variable in levels (regardless of the transformation you did on the data). However, if you use the command: *predict, fcpi xb* Stata will show the forecasts for the transformation. It is all in the manual, in page 3 you have further details. I attached the link for simplicity: www.stata.com/manuals13/tsarimapostestimation.pdf#tsarimapostestimation
Regards, JD
@@JDEconomics Thank you so much, Juan. Your videos and explanation are really very very helpful, thank you very much! Best wishes for the upcoming holidays 🧑🎄
@@dariakravchenko5433 thanks! Merry xmas and happy holidays! JD
Hi thank you for the tutorial. Could you also explain how to forecast using SARIMA model?
Wow Great Explanation ,but i think you must clarify that Eviews shows the inverse roots not the roots itself
Thanks!
AR lags are not significant. Rest all properties work during estimatation. Is that okay to ingore the significance of ar p value
Thank you so much for your video. One question please: because I use the daily data and I set my time with 「gen t =_n」so that my new time series data would be 「t」 and that it starts from 1,2,...,900. How do I type the 「y dynamic(_(__))」about the underline in these two parts？ Thanks
I do have the same doubt. Can u explain what needs to be filled. If u have found out urself.
Hello! Will there be a part 4 for this?
Hi, Thanks for your message. There are only 3 videos, one for each stage of the Box Jenkins Method. What did you want me to cover in video 4?
Regards,
JDEc,
@@JDEconomics I greatly appreciate your immediate response. It would be great if there would be a tutorial on the forecast command and not the predict command. Or are they just practically the same?
Because it seems like different ARIMA models would result to the same predict results. Thank you. Subscribed.
@@naokimikoshiba667 hi! Thanks for your message and subscribing! So you use the predict command to forecast periods ahead. Basically the predict command is used after a model estimation. By typing predict “y” we are telling Stata it has to predict the dependant variable (in our case cpi) and using dynamic we are telling Stata it needs to be a dynamic rather than static prediction. For this example, the graph will look very similar among diverse arima models specifications because as you can see the CPI is an increasing line. Any arima model for this case will fit a line. However if you type “browse” you will see that different arima models produce different predictions numbers. I hope that helped!
Regards, JDEc.
Thank you very much for your valuable videos. Can you please guide me about Multivariate ARIMA models?
I haven't estimated many of those models. I may look into them and make a video some day. Regards, JD
what if p=0 and q=0 ?
how should I proceed
That’s a random walk. There are no significant patterns. You model that as a random walk y= y(-1) +et
Regards
Thanks a lot!!1
You're welcome! Please feel free to subscribe to my channel and share it with your friends/social network. I wish you good luck in your studies/work!
"estat aroots"
invalid subcommand aroots. Can anyone help to solve this problem? Thank you
What’s your stata version?
@@JDEconomics 12.0
Try typing “help estat aroots”
I did and it says no entries found for search on "estat aroots". @@JDEconomics
So do you have any suggestions for that problem? thank you@@JDEconomics