1
00:00:00,390 --> 00:00:04,470
Now it is time to learn about auto regression forecasting methodology.

2
00:00:06,120 --> 00:00:12,930
Keep in mind that this method should be applied to time cities without trend and seasonality.

3
00:00:13,860 --> 00:00:16,440
So if your cities has trend and seasonality.

4
00:00:17,490 --> 00:00:18,080
Remove it.

5
00:00:18,290 --> 00:00:22,640
Using the method stored in the previous sections and then applied this method.

6
00:00:24,090 --> 00:00:26,790
Now, let's understand the concept behind our Doda direction.

7
00:00:28,870 --> 00:00:34,690
Outdoor regression is basically a linear regression model in a linear regression model.

8
00:00:35,020 --> 00:00:40,900
The prediction of output value is based on a linear combination of input values.

9
00:00:41,620 --> 00:00:44,770
This is an equation of a simple linear regression model.

10
00:00:46,020 --> 00:00:53,060
Whitehead is equal to be zero plus BE1 Times X1 hit VI hard.

11
00:00:53,170 --> 00:00:55,510
Is the prediction an X?

12
00:00:55,720 --> 00:01:03,700
Is the input value Beezy to one to be one of the coefficients which are found my optimizing the model

13
00:01:04,360 --> 00:01:05,210
on the training data.

14
00:01:07,240 --> 00:01:17,470
So what we do is this we take a set of X and Y values and feed those to our model, our model then tries

15
00:01:17,470 --> 00:01:24,730
to fit a straight line through these points so that the total value of the errors is minimized.

16
00:01:26,770 --> 00:01:29,650
This gives us a valuable V0 and BE1.

17
00:01:31,650 --> 00:01:35,890
Once we have V0 and B1 values, we can no predictive value.

18
00:01:35,950 --> 00:01:38,580
Why, given the value of X?

19
00:01:41,570 --> 00:01:45,680
We will cover regulation more extensively in the multivariate prediction section.

20
00:01:46,860 --> 00:01:51,980
For now, we need to understand this much, only that for a regression model.

21
00:01:53,040 --> 00:01:56,340
Initially, we will feed a set of X and Y values to our model.

22
00:01:57,600 --> 00:02:04,590
This data is called the training data because using this data, our software will train the model and

23
00:02:04,590 --> 00:02:07,050
find the values of V0 and B1.

24
00:02:09,880 --> 00:02:16,180
Once we have visalo and we win values, we can predict the value of Y for new values of X.

25
00:02:18,170 --> 00:02:23,690
So how do you does in time for these forecasting, if we want to predict the value of some variable,

26
00:02:24,200 --> 00:02:32,810
say, sales for time period plus one, we can use the observation of previous time periods, such as

27
00:02:32,900 --> 00:02:38,930
sales of time period B, B minus one or T minus two and so on.

28
00:02:40,250 --> 00:02:46,460
So if you want to predict the temperature of tomorrow, you can use the temperature of today or yesterday

29
00:02:46,910 --> 00:02:48,670
to build a linear model equation.

30
00:02:51,360 --> 00:03:00,340
And because this regression model uses data from the same input variable at previous times steps, it

31
00:03:00,340 --> 00:03:05,080
is therefore to us auto regression that is regulation of itself.

32
00:03:07,200 --> 00:03:13,720
So order regression model is this only we take the cities usually lag values or bring the model.

33
00:03:14,680 --> 00:03:18,780
Once the model is train, we can predict the future values using it.

34
00:03:20,880 --> 00:03:22,680
Now, let's look at the practical part.

35
00:03:23,190 --> 00:03:26,910
You will get more clarity also by implementing this model in parts of the.
