1
00:00:00,570 --> 00:00:07,950
The next parameter that we can change is a performance function before Dave, let me just undo this

2
00:00:07,950 --> 00:00:10,230
one and back to the member.

3
00:00:11,700 --> 00:00:18,750
Also, you don't need this anymore, since we have not the train function, just run the network one

4
00:00:18,750 --> 00:00:22,920
more time to make sure that we have a good 15 two.

5
00:00:23,330 --> 00:00:25,260
OK, that is good.

6
00:00:25,260 --> 00:00:28,380
And we have a best algorithm for this data.

7
00:00:28,650 --> 00:00:35,280
Now we can move on and change the performance function for changing the performance function.

8
00:00:35,640 --> 00:00:40,020
The one which has been selected as the default is meniscal error.

9
00:00:40,560 --> 00:00:48,550
To check the rest of them, you can just simply type help and neural network and then performance will

10
00:00:48,570 --> 00:00:51,050
be cleared to come out.

11
00:00:51,090 --> 00:00:58,350
We do anticipate here so we can see we have a mean absolute error or performance function.

12
00:00:58,350 --> 00:01:03,920
And if you click on it, you can get more information or just simply to repeat.

13
00:01:03,930 --> 00:01:07,410
And the hell it just gives you the absolute value.

14
00:01:07,830 --> 00:01:11,360
And then the mini square root, which has been set up by default.

15
00:01:11,370 --> 00:01:16,050
We have a sum some of absolute our performance function.

16
00:01:16,410 --> 00:01:24,570
We have SC, which is so squared error, performance function and Tormore again, you can do the experiment,

17
00:01:24,570 --> 00:01:29,220
you can change different performance functions to find the best fit.

18
00:01:29,580 --> 00:01:37,410
But here, for example, Menagh Square Herer is good and actually it's working very well with most of

19
00:01:37,410 --> 00:01:40,020
the systems and most of their network.

20
00:01:40,030 --> 00:01:47,850
So let's just not change it and move on to the next part, which would be choosing the plot function.