Forecasting Methods made simple - Trend Lines

[@adarshyadav1920]

HI could you please help me to know why we used forecast forward for 2 .

[@adarshyadav1920]

In Liner

[@piyushashah1]

Can you please help me to understand what you mean by this Adarsh?

[@kraizyRider]

hello. I want to use months instead of years and its giving me value error(clearly cause its not a number). If i input the months numerically, it will give me wrong predictions. My data (lets suppose) start for May 2013-April 2014 and i want to forecast the member joins for the period against the actual plus for May 2014 and on. i would really appreciate any help.

[@piyushashah1]

Hi Ted, with Excel, you can't use the alphabetical string to represent the month. But, you can use codes for the month. So, instead for May 2013 you can use 1. For June 2013, use 2 and so on. May 2014 would thus be 13. There are other ways to create and use codes, but this is simplest. Did this help?

[@kraizyRider]

Piyush Shah I have tried doing so. I used 5 for may 2013, 6 for June 2013 and so on. the problem is that when i go back to January 2014 where i have to use 1, it will give me a wrong prediction cause X (y=a+bX) is too small and if i use numerical in the form of 1,2,3...,13 my predictions will hit the roof.

[@piyushashah1]

ted January 2014 should be coded as 9. Can you send me your data?

[@rahulkambadur147]

hi, How can you prove that a= 0.006 other than the value that is given in the graph?

[@piyushashah1]

With Excel, you can use solver with non linear optimization. So, ask solver to select A and B to minimize the square deviation between actual and predicted values

[@bachman460]

Is it possible to reload the video? I am unable to access it. Thanks!

[@piyushashah1]

Hi, I checked on a different computer and the video worked. Please try on a different machine and let me know if it works. Per se, KZbin does not allow me to host the video again with the same label.

[@bachman460]

Okay I'll try it out, thanks!

[@ankushzap]

+bachman460 use some different browser.

[@manarafash2168]

Thank you very much for tis video, I have question , why in the first time you used the the Linear trend line wherein the the next time you used the exponential method?.

[@maddio007]

Hello ..Please explain on the value of e taken as 2.718 in exponential trend line example.

[@piyushashah1]

I did not understand the question Madhup. Can you please elaborate? Since the video is only the understand the concept of forecasting, I have not dwelt into details of value of 'e' but have rather used excel formulae to automatically calculate it.

[@MuhammadRiaz-md7qh]

It's a constant value

[@UdayBhargav1]

I have a doubt here why did you use year for linear trendline but code for exponential trendline? If its not compulsory can we use year for both the approaches? And also can you also tell me which is the best approach to be used for different scenarios? And also telling me the difference would be great. Thank you for your kind explanation and for the video.

[@piyushashah1]

It makes no difference in using the code or the year. It gives the same result. When the data has a trend component, you can't use simple exponential smoothing and linear trendlines are better.

[@MuhammadRiaz-md7qh]

Hy Dear Sir kindly made a video on the mixture of time series models same as a series of videos AR.MA, ARMA,ARIMA, SARIMA and Mixture like ARIMA+ANN, SARIMA+ANN

[@piyushashah1]

Thanks for the comments Muhammad. I suggest that you do not use excel for ARIMA and related tools and rather use R or Python.

[@MuhammadRiaz-md7qh]

@@piyushashah1 thank you Sir, can you share with me any link related to R program of arima,sarina,Ann and mixture of these.

[@piyushashah1]

otexts.com/fpp3/

[@MuhammadRiaz-md7qh]

@@piyushashah1 Thank you Sir

[@sunilsalim277]

How to calculate demand planning if we have last 6 months sales data.

[@MuhammadRiaz-md7qh]

Instead of year or quarter same procedure for monthly data.

[@williamhill7312]

Thanks, you really helped my Audit research

[@bangouramohameddoe4277]

HI Piyush, thank for these fast learning video. I'm a student and I have some few questions please. 1) when do we know what is the best method to use to forecast? 2) In the Trend lines method, when or how do realize that we are dealing we a trends in the data? Thank you very much!

[@piyushashah1]

1. Look at the video on methods of forecasting accuracy. I suggest one approach to select the best method 2. The easiest way to get an initial judgement about trend, season, cycle, etc. is to plot the data on a graph

[@bangouramohameddoe4277]

Thank you!

[@piyushashah1]

BANGOURA Mohamed Doe Share the videos with your friends if you have found them useful.