Video 4: Forecasting

Рет қаралды 72,720

Күн бұрын

When we offer point estimates of forecasts, we are incorrectly ignoring uncertainty. This videos shows how to construct confidence intervals for forecasts with a simple linear regression model.
TABLE OF CONTENTS:
00:00 Forecasting
00:10 Two Main Objectives
00:28 Data for Linear Regression Example
01:07 Can we forecast the unknown?
01:26 Regression Result
02:35 Initial Approach to Forecasting
03:50 Sampling = Uncertainty
04:40 A Confidence Interval for the Forecast
05:38 The Standard Error of the Regression
06:14 Constructing the Confidence Interval
07:20 Summary

Пікірлер: 33

@vladfeldfix7825 2 жыл бұрын

This series of 4 videos needs not a LIKE button, but a LOVE button

@avmohanreddy100 4 ай бұрын

You are not understanding how much valuable this truly is. I can't express my gratitude enough, for making me understand this concept. Again, thank you so much.

@Hari-cz3od 6 жыл бұрын

The explanation can't be better! Thank you so so much!

@ishitaahuja6098 5 жыл бұрын

Amazing explanations! Very to the point as well as covering everything!

@user-dz7ix5go9d 5 жыл бұрын

Thank you so much. I wish that our teacher will explain us these topics like you.

@EdmondWanyela 3 жыл бұрын

Wonderful!Thanks for enabling me work on a linear model paper to forecast Air traffic operations in developing countries.

@pds567 5 жыл бұрын

Agreed, best stat video on regression I've ever seen.

@carlosalexandrecosta4234 4 жыл бұрын

Awesome detailed linear regression explanation!!

@jmca_power 5 жыл бұрын

hello, and thank you for making this video, you have helped me prepare one of my mid-term exams

@milliekim5072 5 жыл бұрын

Awesome video! Thank you Sir!

@manaralghamdi4781 2 жыл бұрын

Thank you! Well explained

@kamalhashemi6786 4 ай бұрын

Thank you very much for this series of videos, after watching it I have literary been able to transition from a status of reading my textbook where I could not understand a single idea to a status where every words meant a lot and that is only due to how you explained everything in such a simple way, if someone from INCAE can tell us who is the person making these videos ? thanks !

@hussainmani9439 7 жыл бұрын

best explanation. marvelous

@benjaminchao4742 9 жыл бұрын

Hello! I absolutely love your videos! Do you plan on uploading more videos about advanced statistical analysis?

@engenhariaquimica6590 2 жыл бұрын

Awesome!! Thanks

@yasithaweer Жыл бұрын

Thank you

@yomaru_1999 4 жыл бұрын

very clear

@thefirstShreyas 6 жыл бұрын

Good video. General correction for viewers: 'Subtract' not 'Substract'

@getinettafesetucho7999 9 ай бұрын

Thanks for this marvelous tutorial. I am trying to comprehend this concept. But my Question is this (based on the explanation right on 3:38 minute)! When we interpret the model coefficient, if am not mistaken or someone might correct me, we need to assume that Y-intercept is kept zero, meaning for every 1 unit increase in the explanatory variable (X), there is always X times the model coefficient or coefficient of X. Actually, for the total family consumption (the left of the equation), we need to add the Y-intercept as well. Otherwise, if we add the Y-intercept to it, like in this video case, it will invalidate the meaning our interpretation of the relationship, which means for the family income increment by 100, the consumption will be increased by 85.3. Thank you!

@vkm-hj8ib 7 жыл бұрын

Hi, I really loved the video. I have one question to you. Why we have to consider 2 times of RSE value in finding confidence interval ?

@zlata7264 6 жыл бұрын

because 95% of CI is equal to almost 2 standard deviations (2 STD = 95.45%)

@berhanetedla Жыл бұрын

Appreciated!!!! This professor is erudite. I wonder what he has been working on since this video.

@bhaskarrai5796 3 жыл бұрын

is there any link from where I can get the ppt of the lectures??

@priyankadoiphode5 3 жыл бұрын

Sir please make video on gauss Markov therem in simple regression model 😀😀😀

@abhidcooldude6819 6 жыл бұрын

in 2:35 why we took 0 for error term e????

@zlata7264 6 жыл бұрын

e - is the difference between actual Consumption (Yi) and estimated. In this example, we don't know actual Consumption level of this family (no Yi value), so we try to estimate it = > no error here

@ritzie7611 2 жыл бұрын

Intuitively, if weekly income is 300 and below, then most likely that families under this income bracket category would plunge deeper in debt since their consumption is expected to be greater than their income?

@baotienduong8831 3 жыл бұрын

Hi, why do we have to +/- 2 times the std error of regression instead of 1 time?

@josenino9795 3 жыл бұрын

Because a 95% confidence interval is equal (or close to equal) to 2 standard deviations. So y(x) ± 2*SE will give us the lower and upper of our 95% CI.

@thefirstShreyas 6 жыл бұрын

Shouldn't you use a prediction interval instead of a confidence interval.

@stevenson720 4 жыл бұрын

Lol. NEVER!!!! report a point interval, three circles. And yet newspapers and tv only ever report point intervals, unemployment down by 20,000, well actually its really in the range 20,000 up to 60,000 down, with a 95% certainty.