Dickey-Fuller test for Time Series Stationarity using Python

[@diazjubairy1729]

You are one of the best youtuber for data science in this platform, keep up the good work

[@bhattbhavesh91]

Wow, thanks! I wish KZbin's algorithm understands this :)

[@econtutor10]

Great video! Thanks for the content it is really helpful, only thing to add is that this is not the actual Augmented Dickey Fuller (ADF) test but the Dickey Fuller test (DF).

[@bhattbhavesh91]

My bad

[@valensrwema]

Great work ever , i am very happy for this work , Sir we need more test of time series analysis thank you

[@BhishamSharma-xd3vy]

Statsmodels's ADF has an attribute called "regression' which takes information about trend, no trend, constant etc. And changing the values of this parameter changes the result completely. Can you please discuss about how to choose correct value of "regression" attribute according to any given timeseries?

[@estevaomcs]

great job, king. Keep going!

[@bhattbhavesh91]

Thanks, will do!

[@arthurlucena9898]

Great explanation!!

[@bhattbhavesh91]

Thank you

[@K-mk6pc]

Very good video thanks for Describing but didn't understand one thing why did you Remove the last observation for observation? Claiming it needs to to be equal?

[@unpatel1]

I had the same question, but I think I got it! Say there is a data frame with 5 rows --> 0,1,2,3,4. You are creating diff column with 0',1',2',3' rows. Here row 0' = row 0 - row 1, 1' = 1 - 2, 2' = 2 - 3, 3' = 3 - 4. You could create only 4 rows (0',1',2',3') but the original has 5 rows (0,1,2,3,4). So, you will have to remove the last row (row# 4) from the original data frame.

[@K-mk6pc]

@@unpatel1K man its been very long have to Go throught it

[@uchejonathan6393]

Thanks for the video ,So much to learn. I don't know if it is possible where p-value for the ADF test falls within the three confidence intervals displayed. Please ,if such happens which do I .choose fo enable me compare the corresponding critical value. Thanks

[@khaledmustafa7341]

Hello, I have a question, when I manually find the Dickey Fuller statistic value, the statistic value is very slightly different from the value generated from the Eviews program, although I use the same data, what is the reason?,, I mean the normal Dickey Fuller test, not the developer

[@vibesnovibes6320]

Very intuitive... thanks!

[@bhattbhavesh91]

You're most welcome!

[@prashantb3749]

Thanks for sharing, how we can calculate the ADF TEST values of stationarity for each column say for last 90 readings in a time series. For example - I have stock prices for last 90 hours and going ahead I will be storing the next hours readings in same table. Is there a way to do ADF TEST on rolling basis like we calculate moving average on rolling basis

[@aakashgupta3997]

I thought you will discuss more on the t-stat used in ADF.

[@kingstonxavier]

Is the Dickey Fuller test is different from Augmented Dickey Fuller test? if yes what is the difference?

[@williamgnegonbegonta591]

Sir can i have your mail box? and want to prepare a master research topics using time series analysis or panel date

[@deborahd4991]

Please how many lags were used for this test. I'm trying to test monthly data of 3 years.

[@kujinshin2399]

How could I calculate "Critical Values" without using adfuller function?

[@yashnarendra]

What if the ADF statistic value comes between the critical values. Example - ADF statistics = -2.792; 1% = -3.439; 5% = - 2.866; 10% = -2.569

[@maxwellpatten9227]

excellent video... thanks!

[@bhattbhavesh91]

Glad you liked it!

[@nethrasubash208]

What is exactly ts_values...?

[@chetannare9914]

Also request you to please provide both the data file which you used in above video

[@TheDarshangowda]

Hi do you have the excel sheet? Series 1

[@TheHUPELE]

Hi, good video.Can you also explain math behind it, i mean why LR model between diff of values and actual values verifies as test for stationary

[@vibinthottumgal2685]

Great !Very informative video. Although even if we find the T-stats using a regression model, we need the statsmodel module and adfuller in it to ultimately check the T-stats we found separately to the critical values from adfuller test right? .So in any case the module is necessary. But thanks for explaining how the adfuller test finds the T stats. Could you also make similar videos explaining how the critical values are found so as to reject the null hypothesis. Also it would be very helpful if you could provide an explanation about the third and fourth result values(2nd and 3rd index) in the adf test results(lag and observation) so that I can choose the adequate TS forecast models on the it's basis . Thanks

[@jananiravinag]

phillips perron test - python video if possible please

[@bhattbhavesh91]

Sure :)

[@anubhavsood1510]

Hi Bhavesh, I think there's a typo in your description box relating to the null and alternate hypothesis. You have written the same condition for both.

[@bhattbhavesh91]

Thanks for pointing out! I have made the changes.

[@punniyamoorthyk7382]

nice bro, I have a lots of doubts in data science, please clarify that doubts bro, how do contact you?

[@cooled245]

where is the trend model?

[@chetannare9914]

Hi thanks for video. Can you please provide how can we calculate residuals in a table with all code and data file. It will be a great help.

[@deveshkumar6427]

Time series data should be time dependent . How do we get to know that ??. My data is time dependent or not

[@AkashSingh-mr9ms]

Suppose, I found out that my data is not stationary then can I just directly say that it is seasonal?

[@saqibcs]

P-value formula?

[@muhammadzubairbaloch3224]

Great work. Sir I need the solution of real time data extraction of twitter for crime rate prediction. please make full lecture on this topic. thanks

[@vadimkorontsevich1066]

bro, as not Indian it is VERY difficult to get what you would like to say. It is very difficult to improve pronounciation, hence please write subtitles. It would help a lot of people and make your videos more popular. Good luck, dude.

[@Bharathursharthikote]

Sir can please do video on Logistic regression in python Sir plesae

[@maroben225]

You are good and i love your content but try to add subtitle i mean no offense since you are not a native speaker, it would be really helpful if you can add some subtitles in english. it will really clear the way. but anyway thank you for the content it is a really good chanel we are greatful for the efforts you make sir .

[@Bharathursharthikote]

Thanks

[@sivahanuman4466]

I am getting key error: value