1
00:00:00,780 --> 00:00:09,010
This time, let's create a figure, but we define three different plots and graphs in the same window.

2
00:00:09,660 --> 00:00:15,030
The thing that I'm going to do is divide this one to four parts.

3
00:00:15,390 --> 00:00:18,390
The upper left part would be my feet.

4
00:00:19,080 --> 00:00:22,620
The upper right part would be my regression.

5
00:00:22,920 --> 00:00:31,680
The lower left part shoulder and lower right part showed a histogram for this purpose.

6
00:00:31,680 --> 00:00:39,300
Since we don't have any matlab tool, I need to create a function, open another craib Coleta function,

7
00:00:40,620 --> 00:00:47,700
give it a name like a plot resolves and then in apprentice's, I want to take three different inputs.

8
00:00:48,030 --> 00:00:49,660
One of them is moving target.

9
00:00:49,660 --> 00:00:56,670
The other one is the output of my network and a name which I'm sending this name from the Advanta script.

10
00:00:56,670 --> 00:01:03,840
Number three, if you're creating a function you need to put to end to finish this function here, just

11
00:01:03,840 --> 00:01:12,930
click on top and then start writing your finger and your plots to start with the keyboard figure to

12
00:01:12,930 --> 00:01:13,880
open a figure.

13
00:01:14,610 --> 00:01:20,390
Then here I want to divide my figure into four different parts.

14
00:01:20,940 --> 00:01:29,610
It means I need to use a sub plot and then we have two columns, two rows, and this one is the first

15
00:01:29,610 --> 00:01:29,880
one.

16
00:01:30,900 --> 00:01:36,090
I need to do the same for the third one and fourth one as well.

17
00:01:36,120 --> 00:01:37,410
Here is number two.

18
00:01:37,410 --> 00:01:45,450
The second part, the third part and this one is the forge board gave them so much space to write your

19
00:01:45,450 --> 00:01:46,290
program.

20
00:01:47,610 --> 00:01:58,680
In this part, I will have the feed of targets and my outputs in a plot.

21
00:01:58,680 --> 00:02:02,400
I want to show first the why would a color of black.

22
00:02:04,350 --> 00:02:06,480
We're not done yet with this plot.

23
00:02:06,480 --> 00:02:08,700
So type hold on.

24
00:02:09,540 --> 00:02:12,510
Another plan to show the targets.

25
00:02:12,960 --> 00:02:18,840
We did color off red and I wanted to be dotted dotted line.

26
00:02:21,780 --> 00:02:30,810
And Ålesund with some explanation, like the first one is out Putes, the second one is showing my targets

27
00:02:32,260 --> 00:02:40,170
and of course you'd need to have a title, just have a same name, vete this one that is coming from

28
00:02:40,440 --> 00:02:41,220
advances.

29
00:02:41,250 --> 00:02:41,500
Great.

30
00:02:42,360 --> 00:02:50,540
Now let's move on to the second part, which would be the upper right part of this subplot.

31
00:02:51,120 --> 00:02:53,910
This time I want to show the correlation.

32
00:02:56,020 --> 00:03:04,060
For this purpose, I will plot to target the foods and I will show all of them the color of black,

33
00:03:04,330 --> 00:03:06,430
you can show them weed out.

34
00:03:06,790 --> 00:03:10,720
Usually meth lab is showing them with little black circles.

35
00:03:11,770 --> 00:03:17,260
So just click on or this one would show it with a black little circles.

36
00:03:18,940 --> 00:03:20,480
We are not done yet here.

37
00:03:20,500 --> 00:03:21,340
Hold on.

38
00:03:22,900 --> 00:03:31,250
The next part we need to define X mean and X marks to show where are we going with this data.

39
00:03:31,270 --> 00:03:46,100
So for the X mean, I need to have me off my mean target, which is here t and mean of my neural net

40
00:03:46,130 --> 00:03:50,960
output, which is why for the X marks do the same.

41
00:03:51,430 --> 00:03:55,630
This one is a max of targets in max of output.

42
00:03:56,590 --> 00:03:58,930
And so all of them in a plot.

43
00:03:59,680 --> 00:04:09,670
I will open and close at Brocket inside a bracket I will show the X mean X max and then another bracket

44
00:04:09,670 --> 00:04:20,410
to show the X mean again and X max since we have two inputs and finally show them with the color of

45
00:04:20,410 --> 00:04:33,070
blue and let's show them a little bit t care to do just type line width and maybe give it a number of

46
00:04:33,070 --> 00:04:34,270
two or three.

47
00:04:34,270 --> 00:04:35,630
Two would be good I think.

48
00:04:36,160 --> 00:04:39,580
So this one should be the color of blue and a b t Caroline.

49
00:04:40,180 --> 00:04:43,360
And next point, I'm looking for a correlation.

50
00:04:43,360 --> 00:04:53,400
I will define a variable R and I use the function C or R R Ifill's selected F and F one.

51
00:04:53,440 --> 00:04:56,350
You can see the syntax for this correlation.

52
00:04:56,770 --> 00:05:01,770
I need to put here my target and my outputs of neural network.

53
00:05:02,050 --> 00:05:03,460
But pay attention here.

54
00:05:03,730 --> 00:05:05,380
They need to be in a column.

55
00:05:05,420 --> 00:05:12,090
So I need to use reverse of targets in reverse of outputs to show the correlation.

56
00:05:13,090 --> 00:05:19,690
Finally finished with the title and is titled Let's Call Them.

57
00:05:19,690 --> 00:05:23,680
Correlations are equal to that.

58
00:05:23,680 --> 00:05:26,920
We have a number I want to show it with a string.

59
00:05:26,920 --> 00:05:32,820
So num two s t r which one that are here.

60
00:05:33,430 --> 00:05:34,540
OK, very well.

61
00:05:34,990 --> 00:05:40,660
Now let's move on to the next part and here I want to see the errors.

62
00:05:41,680 --> 00:05:44,370
The first thing to do is calculating an error.

63
00:05:44,530 --> 00:05:50,890
Create a variable name E is equal to the targets minus the outputs.

64
00:05:51,400 --> 00:05:56,020
Finally Shou and should we on a plot give it a color.

65
00:05:56,020 --> 00:05:57,820
Let's say color of blue.

66
00:05:59,710 --> 00:06:08,760
At some guide here, we have ever it can do something else like calculating the mean square error,

67
00:06:09,580 --> 00:06:16,830
create a variable name MCE and this one is the meaning of this obscure error.

68
00:06:20,050 --> 00:06:28,660
We can calculate something else, like a root, mean square error, root, mean square error, we can

69
00:06:29,020 --> 00:06:35,620
define it to be a square off this mean square error very well.

70
00:06:36,880 --> 00:06:39,580
And finally, give it a title.

71
00:06:39,880 --> 00:06:41,680
Again, we have numbers.

72
00:06:41,680 --> 00:06:43,980
We want to convert them to the strings.

73
00:06:44,410 --> 00:06:56,860
So the first part would be mean square error, not to s t r for which one forming a square, another

74
00:06:56,860 --> 00:07:00,370
one for root mean square error.

75
00:07:01,090 --> 00:07:02,540
OK, all set.

76
00:07:03,580 --> 00:07:09,990
Let's move on to the next part and this would be the lower part right side.

77
00:07:10,780 --> 00:07:17,250
This time I want to calculate the histogram for Mr Graham.

78
00:07:17,260 --> 00:07:20,820
We have several functions that we can use this time.

79
00:07:20,860 --> 00:07:33,490
I'm going to use that he's tweaked if you forget about the syntax just selected FMF Flon or just simply

80
00:07:33,490 --> 00:07:35,200
eflornithine some keyboards.

81
00:07:35,650 --> 00:07:38,380
And here I want to give you the 50 being.

82
00:07:38,400 --> 00:07:41,320
So I'm choosing the second syntax.

83
00:07:42,730 --> 00:07:44,170
Show it with a fifty being.

84
00:07:45,250 --> 00:07:53,200
Another thing that we can do is calculate the same for a standard deviation and error and also Minoff

85
00:07:53,200 --> 00:07:53,620
error.

86
00:07:53,800 --> 00:07:58,060
To do that create another variable called E mean.

87
00:07:58,450 --> 00:08:03,010
This time this is equal to the mean of our error.

88
00:08:04,840 --> 00:08:12,770
And here we have another variable for a standard deviation, which I'm going to call it, Steve, the

89
00:08:12,820 --> 00:08:14,110
standard deviation.

90
00:08:14,410 --> 00:08:19,600
And we need to use a function named Usted to calculate that.

91
00:08:20,410 --> 00:08:24,270
Finally, give it a title inside a bracket.

92
00:08:25,000 --> 00:08:28,350
Let me do something a bit interesting.

93
00:08:28,570 --> 00:08:37,870
If you use a backslash and then call the key word Myu, you will be able to see to a room and no new

94
00:08:37,870 --> 00:08:38,740
for this one.

95
00:08:39,220 --> 00:08:45,650
We can do the same for showing a sigma sign in mathematic congregating.

96
00:08:45,660 --> 00:08:56,170
Number two is tweaking another one to show that sigma just use a backslash again and type Sigma do use

97
00:08:56,170 --> 00:08:58,000
the mathematical sine to show.

98
00:08:58,390 --> 00:09:07,570
Just Google it for the codes of Matlab to show different mathematical signs and symbols in a Matlab

99
00:09:07,570 --> 00:09:07,900
code.

100
00:09:09,930 --> 00:09:19,260
OK, I think we are all set to just create a function named plot results Aztek Htwe incudes from the

101
00:09:19,260 --> 00:09:24,160
program that is calling this function, and then it has four different parts.

102
00:09:24,180 --> 00:09:27,060
This is a first part to show defeat of and why.

103
00:09:27,360 --> 00:09:29,940
This is a second part to show Decorrelation.

104
00:09:30,210 --> 00:09:32,460
This is the third part to show the error.

105
00:09:32,470 --> 00:09:35,050
And finally, the last slide shows the histogram.

106
00:09:35,880 --> 00:09:41,550
Let's check the results in advance of Screwtape number three in a plot point.

107
00:09:41,580 --> 00:09:43,180
I don't need them more.

108
00:09:44,290 --> 00:09:49,710
Remove them and call for the plot results.

109
00:09:50,220 --> 00:09:54,480
Give it a target, which is t the output, which is why.

110
00:09:54,480 --> 00:09:58,470
And give you the name, let's call it all data.

111
00:09:59,310 --> 00:10:00,290
Run the program.

112
00:10:01,020 --> 00:10:02,040
Oh, we have it.

113
00:10:02,710 --> 00:10:09,300
OK, we we missed something we didn't save our flight control is to save the fight.

114
00:10:09,930 --> 00:10:12,000
Very important reminder here.

115
00:10:12,010 --> 00:10:20,430
Make sure you are saving dysfunction four in a same folder with your program, otherwise you cannot

116
00:10:20,430 --> 00:10:24,060
call it and you will get a serious problem plot result.

117
00:10:24,060 --> 00:10:28,080
This is a name of my function and say within a same folder.

118
00:10:29,320 --> 00:10:37,540
OK, here we can see our function now, it should walk back here and click on Ron very well.

119
00:10:38,050 --> 00:10:41,980
We can see all data here in this graph.

120
00:10:42,010 --> 00:10:49,960
We can see the correlation, these little black circles that we created to show our data.

121
00:10:49,960 --> 00:10:52,520
And we have 1000 something of them.

122
00:10:53,230 --> 00:10:55,270
This is my correlation.

123
00:10:55,690 --> 00:11:02,140
The square root is three hundred forty six, the root of minus square, which is showing the average.

124
00:11:02,140 --> 00:11:03,810
And that's more accurate.

125
00:11:03,820 --> 00:11:05,600
That's actually more useful.

126
00:11:05,830 --> 00:11:07,060
It's around 18.

127
00:11:07,390 --> 00:11:08,350
Look at my data.

128
00:11:08,350 --> 00:11:12,880
My data is from minus one hundred two positive one hundred.

129
00:11:13,150 --> 00:11:14,930
But the root is square mirror.

130
00:11:14,980 --> 00:11:17,360
It means most of the data are here.

131
00:11:17,620 --> 00:11:18,630
That is correct.

132
00:11:18,640 --> 00:11:26,070
And it must be actually close to the number of sigma, which is already close.

133
00:11:26,110 --> 00:11:28,600
It's not the same, but it's very close.

134
00:11:29,500 --> 00:11:36,700
This one is the most sign which we created with the Matlab code and it shows you and this one is our

135
00:11:36,700 --> 00:11:37,270
sigma.

136
00:11:37,450 --> 00:11:39,550
The histogram is also a good fit.

137
00:11:39,550 --> 00:11:47,890
It's around zero and you can do the same for your training data, for your validation data and for your

138
00:11:47,890 --> 00:11:50,930
test data to do that.

139
00:11:51,250 --> 00:11:55,170
Let me just show you one of them here.

140
00:11:55,180 --> 00:11:58,780
We use the mouse to separate the train data.

141
00:11:58,930 --> 00:12:04,540
The validation data and to set data seems to have more than one in food.

142
00:12:04,540 --> 00:12:08,510
We need to define it for a specific input.

143
00:12:08,800 --> 00:12:14,650
This time I'm going to choose the first input, just copy pasted for the rest of them.

144
00:12:14,660 --> 00:12:17,340
Otherwise you will get a logical error.

145
00:12:18,910 --> 00:12:30,300
Now, from the first input, I want to see this one for the train target and train outputs to do that.

146
00:12:30,310 --> 00:12:33,610
Let's just copy paste it one more time.

147
00:12:33,970 --> 00:12:37,780
This time this is my train target.

148
00:12:38,620 --> 00:12:42,380
Is that the target or target targets.

149
00:12:42,790 --> 00:12:43,100
Good.

150
00:12:43,240 --> 00:12:46,570
And this one is my train.

151
00:12:48,490 --> 00:12:54,570
Output's the name would be train date.

152
00:12:55,380 --> 00:12:59,240
You can do the same for validation and for testing as well.

153
00:13:04,520 --> 00:13:12,130
OK, here it is, I already created for validation data, which is valid targets and the name for Alby's

154
00:13:12,160 --> 00:13:13,370
Isma'il output.

155
00:13:13,640 --> 00:13:15,500
I just changed a name as well.

156
00:13:15,710 --> 00:13:17,190
I did the same for test.

157
00:13:17,210 --> 00:13:18,780
Now let's run your program.

158
00:13:19,160 --> 00:13:21,430
We need the prentice's here.

159
00:13:21,920 --> 00:13:25,320
But other than that, it should work.

160
00:13:26,150 --> 00:13:27,500
OK, let's run.

161
00:13:30,230 --> 00:13:32,810
Well, we have four different figures.

162
00:13:33,080 --> 00:13:35,240
The first one is for all data.

163
00:13:35,930 --> 00:13:37,910
This one is for the train data.

164
00:13:37,940 --> 00:13:42,200
We can see the histogram correlation, all the things that we are looking for.

165
00:13:42,200 --> 00:13:51,020
If you need more, you can just add and have a subplot just divided more maybe to support two and three

166
00:13:51,020 --> 00:13:52,490
to have more columns.

167
00:13:53,090 --> 00:13:56,040
This was all for the first output.

168
00:13:56,510 --> 00:14:01,700
Let's see what can we do if we want to see for the second output?