Time-Series Analysis with R | 1. Decomposition

Поделиться
HTML-код
  • Опубликовано: 10 дек 2024

Комментарии • 78

  • @zihaoliang9208
    @zihaoliang9208 4 года назад +3

    I'm majoring in SQM and your videos are all amazing, much cleaner and easier to understand than my prof's lecture. I'm taking my time to go through every single video on this channel. Thank you and a big thumbs up.

    • @bkrai
      @bkrai  4 года назад

      Great to hear!

  • @koushalrayi
    @koushalrayi 5 лет назад +4

    Thank you very much sir. This is so straight-forward, clutter-free from excess info and easy to understand. It would be great if somewhere you mention about some specifics like lag.max and what values should be taken and why? , etc..

    • @bkrai
      @bkrai  5 лет назад

      Thanks for the feedback!

  • @skfunnext
    @skfunnext 5 лет назад +2

    Thanks sir for such a simple way to explain the concept! Please include some real data as well in next video of series which helps to understand the real world issues comfortably

    • @bkrai
      @bkrai  5 лет назад +3

      Thanks for comments and suggestion! For these introductory time-series I'll still use simpler examples keeping in mind many may be doing it for the first time. But later I plan to work on more complex situations too.

  • @flamboyantperson5936
    @flamboyantperson5936 5 лет назад +2

    Great tutorial Sir. Very informative and nice. Thank you Sir.

  • @redarabie7098
    @redarabie7098 5 лет назад +2

    Great tutorial Sir. very useful for us thank you very much Sir

    • @bkrai
      @bkrai  5 лет назад

      Thanks for comments!

  • @muhammadazkas2738
    @muhammadazkas2738 2 года назад +1

    sir, what is the meaning of "type" and "las" when u write the code of plotting

  • @mohamedfarag9584
    @mohamedfarag9584 Год назад

    Thanks a lot for your simplified explanation. Could you please help us through a lecture explaining date formatting before converting the data frames into data series?

  • @earlymorningcodes6100
    @earlymorningcodes6100 4 года назад +1

    4:18 Decomposition

  • @al38261
    @al38261 Год назад +1

    Thank you very much!

    • @bkrai
      @bkrai  Год назад

      You're welcome!

  • @liwenling1287
    @liwenling1287 2 года назад +1

    Hi Professor Rai, thanks for your video on time series. Here I have a question on the 'ts' command. If my data is a daily one, how should i set the frequency? how can i specify the particular starting/ending date?

  • @smsm314
    @smsm314 5 лет назад +1

    Good Morning professor,
    Please professor,
    1)- if we have the Yt series (with trend). To study the stationarity of this series, we can do the following decomposition (or filtering):
    Yt = F(t) + Ut, such that F(t) is stationary at the turn of deterministic trend (that is, without unit root) (linear, nonlinear (Quadratical, Cubic...). And if we find the series Ut it is stationary, it implies that Yt it does not have a unit root, and the opposite is right?
    2)- If a series it is Ts (Trend Stationary) at the level, it imply that this series it is I (0) and not I(1)?
    Best wishes

  • @ShubhamKumar-xy6kj
    @ShubhamKumar-xy6kj 4 года назад +1

    Sir, can we sample from known standard probability models of time series.
    Just I wanted to observe the ACVF plot for different processes.

    • @bkrai
      @bkrai  4 года назад

      Try and see what you get.

    • @ShubhamKumar-xy6kj
      @ShubhamKumar-xy6kj 4 года назад

      Then how to sample from such processes?

  • @movie4dilip
    @movie4dilip 4 года назад +1

    Pretty nice, thanks

    • @bkrai
      @bkrai  4 года назад

      You are welcome!

  • @earlymorningcodes6100
    @earlymorningcodes6100 4 года назад +1

    3:46 log Transformation

  • @tripplannermumbai4499
    @tripplannermumbai4499 4 года назад +1

    decomp

  • @ravikumarut
    @ravikumarut 5 лет назад +2

    Sir , is there any video on Oracle log file analysis anomoly detaction using R and ML !

    • @bkrai
      @bkrai  5 лет назад

      I do not have one at this time.

  • @mdsaifulislam5891
    @mdsaifulislam5891 5 лет назад +1

    dataset = read.csv('airline_passengers.csv')
    > class(dataset)
    "data.frame"
    > start(dataset)
    Error in attr(x, "tsp")

    • @bkrai
      @bkrai  5 лет назад +1

      If the data has a header, you may have to use header = T while reading csv file.

    • @mdsaifulislam5891
      @mdsaifulislam5891 5 лет назад

      @@bkrai Sir thank you. I will try it

  • @kenn756
    @kenn756 4 года назад +1

    Is this split basis the additive or the multiplicative model?

    • @bkrai
      @bkrai  4 года назад +1

      It is additive.

    • @kenn756
      @kenn756 4 года назад

      @@bkrai is there any package in R that deals with the multiplicative model?

  • @ssetyawan
    @ssetyawan 4 года назад +1

    Thank you for your valuable lecture sir,
    I want to ask :
    after I load the data by ' data(AirPassenger )' and convert to AP object. but when I want to know the Header column with function ' head(AP)' why it shown =
    AP head(AP)
    [1] 112 118 132 129 121 135 #the header was not shown like your's

    • @bkrai
      @bkrai  4 года назад

      Are the next steps working correctly?

    • @ssetyawan
      @ssetyawan 4 года назад +1

      Yes, no problem. All next step is working properly. Just issue on the heading of the data

    • @bkrai
      @bkrai  4 года назад

      Ok, thanks for the update!

  • @ukpraveen
    @ukpraveen 5 лет назад +1

    Hi Rai, very good videos really enjoying it, please can you tell where to get data to practice

    • @bkrai
      @bkrai  5 лет назад +1

      This data is in R itself. You can access it by using the code that I used.

  • @facehappyable
    @facehappyable 4 года назад +1

    Thanks for the videos. I want to know how to compare two time-series data. I am a geologist, and want to compare the time series of the same variable (such as temperature) at two locations.

    • @bkrai
      @bkrai  4 года назад

      You can plot both on the same graph.

  • @shuvhamdigitalacademy3228
    @shuvhamdigitalacademy3228 3 года назад +1

    Sir, how to make an ordinary dataset to time-series data?
    I have 10 Rows with 10 variables and 29 observations.
    1st row and column is year.
    2nd is Agriculture Employment.
    3rd is Industry Employment.
    4th is Service Sector Emp.
    5th is education.
    6th is Life expectancy.
    7th is population growth.
    8th is GDP growth.
    9th is Inflation rate.
    And 10th is Interest rate.
    Three dependent variable and 6 independent variables.

    • @bkrai
      @bkrai  3 года назад

      For time series you need at least 2-3 years of data to capture trends and seasonality.

  • @harishnagpal21
    @harishnagpal21 4 года назад +1

    Hi Bhartendra, thanks for the video. One query, what if the data is daily for example COVID-19 data. What frequency we should take. Can we take 365?

    • @bkrai
      @bkrai  4 года назад +1

      Yes, try that.

    • @harishnagpal21
      @harishnagpal21 4 года назад

      Thanks. I tried that and it worked.

  • @lianjek5788
    @lianjek5788 4 года назад +1

    Hi Sir, i have monthly ts data that has 216 observations with decimals. I removed those decimals. Then when I bring it in r and try to plot it its coming different and clumsy. Further wehen I add another data series "year" in to the data file, the figure is coming as dot. So the trend is clear but the seasonal cycle of the figure is not coming, the way you have done in your slides. In the decomposition I found another error:
    Error in decompose(d) : time series has no or less than 2 periods
    Would you please suggest me what is the problem in my analysis, Do I need to install any package?
    Here are my r codes:
    d

    • @bkrai
      @bkrai  4 года назад +1

      To capture seasonality make sure you have data for 2 or more periods. Currently it detects "no or less than 2 periods".

    • @lianjek5788
      @lianjek5788 4 года назад +1

      @@bkrai hi sir, many thanks for the reply. how to capture seasonality… My data is from Dec 2001 to December 2019 and already deleted Dec 2001 data. Therefore I tried monthly, quarterly...so on...
      ts(d, frequency = 12, start=c(2002,1))
      ts(d, frequency = 4, start=c(2002,1:3))
      ts(d, frequency = 3, start=c(2002,1:4))
      Its coming the same error!

    • @bkrai
      @bkrai  4 года назад +1

      Probably you can try this:
      ruclips.net/video/rV-hhKBRKbI/видео.html

  • @tuhinsuryachakraborty
    @tuhinsuryachakraborty 5 месяцев назад +1

    Thanks sir

    • @bkrai
      @bkrai  5 месяцев назад

      You are welcome!

  • @kashifjavedlone1780
    @kashifjavedlone1780 4 года назад +1

    What us the benefit of taking log?

    • @bkrai
      @bkrai  4 года назад +1

      If you look at before and after log transformation plots, it will show how it changes from non-stationary to a much better pattern for prediction purposes.

  • @chethanbk7747
    @chethanbk7747 5 лет назад +1

    sir, am learning predictive subject.. in that am finding too much tough in Time series.. can u suggest me some links to start with time series

    • @bkrai
      @bkrai  5 лет назад +1

      You can try this playlist:
      ruclips.net/p/PL34t5iLfZddt9X6Q6aq0H38gn-_JQ1RjS

  • @gauravsarin4725
    @gauravsarin4725 5 лет назад +1

    Pls share the R file.

    • @bkrai
      @bkrai  5 лет назад

      Link to R file is in description area of following video:
      ruclips.net/video/VQ-9BzKYVIo/видео.html

  • @mdsaifulislam5891
    @mdsaifulislam5891 5 лет назад

    Dear Sir, Can you please give the data set & code. Thanks and regards

    • @bkrai
      @bkrai  5 лет назад +1

      The data is already available with R. If you start with 1st video in the playlist, you can see how to access it:
      ruclips.net/p/PL34t5iLfZddt9X6Q6aq0H38gn-_JQ1RjS

    • @bkrai
      @bkrai  5 лет назад +1

      I'll try to find the code.

  • @DnyaneshwarPanchaldsp
    @DnyaneshwarPanchaldsp 3 года назад +1

    👌👌💐💐

    • @bkrai
      @bkrai  3 года назад

      Thanks!

    • @DnyaneshwarPanchaldsp
      @DnyaneshwarPanchaldsp 3 года назад

      @@bkrai सर I request you आप aspect based sentiment analysis पे video बनाओ...... Plz

  • @kaswin6527
    @kaswin6527 5 лет назад +1

    I wanna be your student for R ..
    How can I ?

    • @bkrai
      @bkrai  5 лет назад +4

      If you are watching my videos, your are already my student :)

    • @kaswin6527
      @kaswin6527 5 лет назад +1

      @@bkrai thanks so much. :)

    • @bkrai
      @bkrai  5 лет назад

      welcome!

    • @mdsaifulislam5891
      @mdsaifulislam5891 5 лет назад +1

      @@bkrai you are the most amazing person who always share his knowledge without anything.

    • @bkrai
      @bkrai  5 лет назад

      Thanks for comments!