wow! Sir, Please continue making videos your videos are very high quality. Anyone can tell by watching you videos that this guys know stuff and your Coding style is Pro
Simply wow Srivatsan.. great content with good clarity of explanation.. even a school kid can understand your content sir with min knowledge.. Thank you soo much.
I have a small comment (more of an errata). At time-stamp of 8:39, in this video you define the formula for Moving Average. I guess the denominator should be (n+1) instead of n.
Subhasis.. it is typically number of period n. You can check details here en.m.wikipedia.org/wiki/Moving_average#Simple_moving_average Said that there is no harm to do n+1 if data is volatile but in that case I might go for weighted averages
@@AIEngineeringLife ..Well! the denominator should depict the number of observations, which is what the definition of average entails. This is consistent in the link you shared too. Now in the video, you have observations at t, t-1, t-2,..,t-n. This is a total of (n+1) observations and thus, my comment that the denominator should be (n+1) and not 'n'.
@@subhasismishra5001 you are right. Problem was in the way I depicted my numerator. It must be ending with n-1 since I started with t0. Thanks for pointing it 👍
@@AIEngineeringLife YOu have done a fabulos job in putting so much effort in these modules. And I am sure most of us are grateful to you for the same. My intention was not to point out the error but just to put it as an errata. Did try to contact you over linkedin, but I was not able to ... Again, thanks for the efforts
@@subhasismishra5001 Thank you and no issues on pointing out as mistakes from my end might happen. It is good for one seeing the comments to know. Maybe I typed the formula in a hurry and did not look into details
Excellent intro to a very basic time series algo. Hope you also cover methods like regression, LSTM and multi-step forecasting, also multiple sku forecasting etc.... Thanks for this video :)
@@AIEngineeringLife please upload videos on daily basis so that we can get more knowledge this is just request from my side doing a great job to DS community!!!!!!!!!!!
Neha.. If you see I am doing forecast for one time period. Now you can do recursive forecasting by taking the last forecasted value along with history and do next value and so on. ewm are good for short term forecasts where we might not except much sudden trend reversal
Thank you sir for a great explanation with practicals. It will be great and easy for future doubt clearance if you provide the links for the dataset and Notebook in the video description.
Great videos. Really enjoy going through these videos. A lot of the concepts from pandas and Matplotlib got revisited. However, I have a suggestion if you could post the links of the data in the description it would be a bit easy for the learners to access it. Thank You for the video. :)
Sure will do it going forward. Most of the datasets are in my git folder here - github.com/srivatsan88/UA-camLI/tree/master/dataset Thank you and glad you liked the videos
Hello sir, I am really gaining so much knowledge from your videos, and thank you so much for making such an amazing channel. Just wondering for the 5 minutes(for a year) data what can be the suggestion for the moving average.
If u keep it constant it will be like SMA but u will end up under weighting output if it does not equal to n. If it equals to n then it will be 1 and exactly SMA
for learning purpose i recreated the notebook and as i go down the playlist, i will as well gradually update the code. github.com/mbjallow6/data-science-projects
Really good series about time series analysis. He knows deep about what he is teaching.
Hello Srivats. These videos are very lucid and simple. I am learning a lot from them. Thanks for taking the time and effort for doing so.
wow! Sir, Please continue making videos your videos are very high quality. Anyone can tell by watching you videos that this guys know stuff and your Coding style is Pro
Very thoroughly explained.
Simply wow Srivatsan.. great content with good clarity of explanation.. even a school kid can understand your content sir with min knowledge.. Thank you soo much.
I have a small comment (more of an errata). At time-stamp of 8:39, in this video you define the formula for Moving Average. I guess the denominator should be (n+1) instead of n.
Subhasis.. it is typically number of period n. You can check details here
en.m.wikipedia.org/wiki/Moving_average#Simple_moving_average
Said that there is no harm to do n+1 if data is volatile but in that case I might go for weighted averages
@@AIEngineeringLife ..Well! the denominator should depict the number of observations, which is what the definition of average entails. This is consistent in the link you shared too. Now in the video, you have observations at t, t-1, t-2,..,t-n. This is a total of (n+1) observations and thus, my comment that the denominator should be (n+1) and not 'n'.
@@subhasismishra5001 you are right. Problem was in the way I depicted my numerator. It must be ending with n-1 since I started with t0. Thanks for pointing it 👍
@@AIEngineeringLife YOu have done a fabulos job in putting so much effort in these modules. And I am sure most of us are grateful to you for the same. My intention was not to point out the error but just to put it as an errata. Did try to contact you over linkedin, but I was not able to ...
Again, thanks for the efforts
@@subhasismishra5001 Thank you and no issues on pointing out as mistakes from my end might happen. It is good for one seeing the comments to know. Maybe I typed the formula in a hurry and did not look into details
Excellent intro to a very basic time series algo. Hope you also cover methods like regression, LSTM and multi-step forecasting, also multiple sku forecasting etc....
Thanks for this video :)
Yes I do have plans on advanced model. Initially covering the foundational concepts and will follow other algos post that
Thanks and please continue the good work 👍
@@AIEngineeringLife please upload videos on daily basis so that we can get more knowledge this is just request from my side doing a great job to DS community!!!!!!!!!!!
Thank you Srivatsan sir.. Its really helpful..😊
How do I make future forecast using EWM or other moving average techniques. Also Thank you sir for your thorough explanation
Neha.. If you see I am doing forecast for one time period. Now you can do recursive forecasting by taking the last forecasted value along with history and do next value and so on. ewm are good for short term forecasts where we might not except much sudden trend reversal
Awesome content.
Thank you sir for a great explanation with practicals.
It will be great and easy for future doubt clearance if you provide the links for the dataset and Notebook in the video description.
Thanks @Sri. Very elaborative.
Thanks Sarath
Great videos. Really enjoy going through these videos. A lot of the concepts from pandas and Matplotlib got revisited.
However, I have a suggestion if you could post the links of the data in the description it would be a bit easy for the learners to access it.
Thank You for the video. :)
Sure will do it going forward. Most of the datasets are in my git folder here - github.com/srivatsan88/UA-camLI/tree/master/dataset
Thank you and glad you liked the videos
Amazing explanation
Hello sir, I am really gaining so much knowledge from your videos, and thank you so much for making such an amazing channel. Just wondering for the 5 minutes(for a year) data what can be the suggestion for the moving average.
nice explanation sir could you upload all the code in your git hub which will be useful for all to practice and get some ideas.
Here you go - github.com/srivatsan88/End-to-End-Time-Series/blob/master/Moving_Averages.ipynb
What is the difference between moving averages methonds and smothing methods
In smoothing methods you are just controlling the variations between time steps whereas MAs treat every instance as same
Simple moving average is just a Weighted moving average where weight = 1, what will happen if we keep the weight constant in WMA ?
If u keep it constant it will be like SMA but u will end up under weighting output if it does not equal to n. If it equals to n then it will be 1 and exactly SMA
Adam optimizer ,Rms prop uses this concept
for learning purpose i recreated the notebook and as i go down the playlist, i will as well gradually update the code. github.com/mbjallow6/data-science-projects
This is amazing. All the best