Abstract image of a woman sitting on books and studying on a laptop

4 Dimensionality reduction (PCA & t-SNE)

Dmytro Fishman, profile picture
Dmytro Fishman

The document introduces the concept of dimensionality reduction in machine learning, focusing on visualizing multi-dimensional instances in a two-dimensional space. It discusses the importance and techniques of reducing dimensionality to facilitate data analysis and improve model performance. The document includes several visual representations and examples to illustrate these concepts.

Education
Related topics:
Principal Component Analysis
1 of 93
Downloaded 101 times
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
Most read
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
Most read
48
49
50
51
52
53
54
55
56
57
58
59
60
Most read
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
Introduction to Machine Learning (Dimensionality reduction) Dmytro Fishman (dmytro@ut.ee)
So far we have been fearless drawing multi-dimensional instances on two-dimensional planes
28px 28px
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 126, 213, 34, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 19, 227, 254, 84, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 114, 254, 254, 84, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 178, 254, 254, 84, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 60, 236, 254, 254, 84, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 78, 254, 254, 254, 84, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 78, 254, 254, 231, 56, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 191, 254, 254, 185, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 233, 254, 254, 105, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 233, 254, 254, 30, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 93, 248, 254, 243, 27, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 16, 192, 254, 254, 130, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 32, 254, 254, 248, 95, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 41, 254, 254, 231, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 186, 254, 254, 160, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 21, 203, 254, 188, 12, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 86, 254, 255, 179, 8, 4, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 86, 254, 254, 254, 254, 130, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 68, 240, 254, 254, 250, 103, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 92, 255, 207, 91, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0 28px 28px
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 126, 213, 34, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 19, 227, 254, 84, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 114, 254, 254, 84, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 178, 254, 254, 84, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 60, 236, 254, 254, 84, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 78, 254, 254, 254, 84, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 78, 254, 254, 231, 56, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 191, 254, 254, 185, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 233, 254, 254, 105, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 233, 254, 254, 30, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 93, 248, 254, 243, 27, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 16, 192, 254, 254, 130, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 32, 254, 254, 248, 95, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 41, 254, 254, 231, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 186, 254, 254, 160, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 21, 203, 254, 188, 12, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 86, 254, 255, 179, 8, 4, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 86, 254, 254, 254, 254, 130, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 68, 240, 254, 254, 250, 103, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 92, 255, 207, 91, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0 28px 28px 102 41
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 126, 213, 34, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 19, 227, 254, 84, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 114, 254, 254, 84, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 178, 254, 254, 84, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 60, 236, 254, 254, 84, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 78, 254, 254, 254, 84, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 78, 254, 254, 231, 56, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 191, 254, 254, 185, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 233, 254, 254, 105, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 233, 254, 254, 30, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 93, 248, 254, 243, 27, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 16, 192, 254, 254, 130, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 32, 254, 254, 248, 95, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 41, 254, 254, 231, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 186, 254, 254, 160, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 21, 203, 254, 188, 12, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 86, 254, 255, 179, 8, 4, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 86, 254, 254, 254, 254, 130, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 68, 240, 254, 254, 250, 103, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 92, 255, 207, 91, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0 28px 28px 41 Pixel #X Pixel#Y 102
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 126, 213, 34, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 19, 227, 254, 84, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 114, 254, 254, 84, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 178, 254, 254, 84, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 60, 236, 254, 254, 84, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 78, 254, 254, 254, 84, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 78, 254, 254, 231, 56, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 191, 254, 254, 185, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 233, 254, 254, 105, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 233, 254, 254, 30, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 93, 248, 254, 243, 27, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 16, 192, 254, 254, 130, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 32, 254, 254, 248, 95, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 41, 254, 254, 231, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 186, 254, 254, 160, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 21, 203, 254, 188, 12, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 86, 254, 255, 179, 8, 4, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 86, 254, 254, 254, 254, 130, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 68, 240, 254, 254, 250, 103, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 92, 255, 207, 91, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0 28px 28px Pixel #X Pixel#Y 41 How about the rest of 782 pixels? 102
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 126, 213, 34, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 19, 227, 254, 84, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 114, 254, 254, 84, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 178, 254, 254, 84, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 60, 236, 254, 254, 84, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 78, 254, 254, 254, 84, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 78, 254, 254, 231, 56, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 191, 254, 254, 185, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 233, 254, 254, 105, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 233, 254, 254, 30, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 93, 248, 254, 243, 27, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 16, 192, 254, 254, 130, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 32, 254, 254, 248, 95, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 41, 254, 254, 231, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 186, 254, 254, 160, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 21, 203, 254, 188, 12, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 86, 254, 255, 179, 8, 4, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 86, 254, 254, 254, 254, 130, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 68, 240, 254, 254, 250, 103, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 92, 255, 207, 91, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0 28px 28px Pixel #X Pixel#Y 41 How about the rest of 782 pixels? But we cannot really visualise all 784 dimensions... 102
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 126, 213, 34, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 19, 227, 254, 84, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 114, 254, 254, 84, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 178, 254, 254, 84, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 60, 236, 254, 254, 84, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 78, 254, 254, 254, 84, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 78, 254, 254, 231, 56, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 191, 254, 254, 185, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 233, 254, 254, 105, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 233, 254, 254, 30, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 93, 248, 254, 243, 27, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 16, 192, 254, 254, 130, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 32, 254, 254, 248, 95, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 41, 254, 254, 231, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 186, 254, 254, 160, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 21, 203, 254, 188, 12, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 86, 254, 255, 179, 8, 4, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 86, 254, 254, 254, 254, 130, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 68, 240, 254, 254, 250, 103, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 92, 255, 207, 91, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0 28px 28px Pixel #X Pixel#Y 41 How about the rest of 782 pixels? 102 But we cannot really visualise all 784 dimensions... or can we?
Dimensionality reduction Patients Healthy
The following material was inspired by: https://www.youtube.com/watch?v=_UVHneBUBW0
The following material was inspired by: https://www.youtube.com/watch?v=_UVHneBUBW0 With a bit of fixes, though
Patients Healthy
Patients Healthy Each dot in this plot represents a person
Patients Healthy Every person in its turn could be described by hundreds of protein reactivities Each dot in this plot represents a person
Patients Healthy Every person in its turn could be described by hundreds of protein reactivities Each dot in this plot represents a person The idea is that people with similar protein reactivity profiles cluster together
Patients Healthy 5, 98, 43, …, 9, 66, Patient
Patients Healthy 5, 98, 43, …, 9, 66, Patient 62, 8, 14, …, 73, 2, Healthy
Patients Healthy 5, 98, 43, …, 9, 66, Patient 62, 8, 14, …, 73, 2, Healthy 3, 45, 13, …, 7, 54, Patient
Patients Healthy 5, 98, 43, …, 9, 66, Patient 62, 8, 14, …, 73, 2, Healthy 3, 45, 13, …, 7, 54, Patient How is that even possible?
Patients Healthy PCA 5, 98, 43, …, 9, 66, Patient 62, 8, 14, …, 73, 2, Healthy 3, 45, 13, …, 7, 54, Patient How is that even possible?
Introduction to dimensions
1-Dimensional data Person Protein #1 A 24 B 63 C 51 D 34 E 15
1-Dimensional data 300 60 Person Protein #1 A 24 B 63 C 51 D 34 E 15
1-Dimensional data 300 60 300 60 Uniformly distributed data Person Protein #1 A 24 B 63 C 51 D 34 E 15
1-Dimensional data 300 60 300 60 Uniformly distributed data 30 60 Non-uniform distribution 0 Person Protein #1 A 24 B 63 C 51 D 34 E 15
2-Dimensional data 300 60 Protein #1 Person Protein #1 A 24 B 63 C 51 D 34 E 15
2-Dimensional data 30 60 Protein#2 Person Protein #1 Protein #2 A 24 B 63 C 51 D 34 E 15 300 60 Protein #1 0
2-Dimensional data Person Protein #1 Protein #2 A 24 29 B 63 C 51 D 34 E 15 A30 60 Protein#2 300 60 Protein #1 0
2-Dimensional data A Person Protein #1 Protein #2 A 24 29 B 63 C 51 D 34 E 15 30 60 Protein#2 300 60 Protein #1 0
2-Dimensional data A Person Protein #1 Protein #2 A 24 29 B 63 59 C 51 D 34 E 15 B 30 60 Protein#2 300 60 Protein #1 0
2-Dimensional data A B And so on… Person Protein #1 Protein #2 A 24 29 B 63 59 C 51 D 34 E 15 30 60 Protein#2 300 60 Protein #1 0
2-Dimensional data Person Protein #1 Protein #2 A 24 29 B 63 59 C 51 32 D 34 56 E 15 4 … … … 30 60 Protein#2 300 60 Protein #1 0
2-Dimensional data Protein profiles are correlated Person Protein #1 Protein #2 A 24 29 B 63 59 C 51 32 D 34 56 E 15 4 … … … 30 60 Protein#2 300 60 Protein #1 0
2-Dimensional data Here is no correlation Person Protein #1 Protein #2 A 24 29 B 63 59 C 51 32 D 34 56 E 15 4 … … … 30 60 Protein#2 300 60 Protein #1 0
3-Dimensional data Person Protein #1 Protein #2 A 24 29 B 63 59 C 51 32 D 34 56 E 15 4 … … … 30 60 Protein#2 300 60 Protein #1 0
3-Dimensional data Person P1 P2 P3 A 24 29 B 63 59 C 51 32 D 34 56 E 15 4 … … … Protein #330 60 Protein#2 300 60 Protein #1 0 30 60
3-Dimensional data Person P1 P2 P3 A 24 29 33 B 63 59 C 51 32 D 34 56 E 15 4 … … … A 30 60 Protein #330 60 Protein#2 300 60 Protein #1 0
Dimensions 300 60 1-Dimensional
Dimensions 300 60 30 6 3 6 2-Dimensional 1-Dimensional
Dimensions 300 60 30 6 3 6 3-Dimensional 30 6 3 6 3 6 2-Dimensional 1-Dimensional
How about 200-D data?
Are all D are equally important? How about 200-D data?
Are all D are equally important? Let’s go back to 2-D example How about 200-D data?
30 60 Protein#2 300 60 Protein #1 0 Importance of axes
Variation is from left to right 30 60 Protein#2 300 60 Protein #1 0 Importance of axes
Projected data - does not look too different 30 60 Protein#2 300 60 Protein #1 0 Importance of axes
We can safely remove 2nd dimension 30 60 Protein#2 300 60 Protein #1 0 300 60 Protein #1 Importance of axes Projected data - does not look too different
30 60 Protein#2 300 60 Protein #1 0 300 60 Protein #1 The important variation is from left to right We can safely remove 2nd dimension Importance of axes Projected data - does not look too different
30 60 Protein#2 300 60 Protein #1 0 Principle components
30 60 Protein#2 300 60 Protein #1 0 Data is mostly spread along this line Principle components
30 60 Protein#2 300 60 Protein #1 0 Data is mostly spread along this line and a little bit along this line Principle components
30 60 Protein#2 300 60 Protein #1 0 Data is mostly spread along this line and a little bit along this line What if we make new axes from these lines? Principle components
30 60 Protein #2 30 0 60 Protein #1 0 New X and Y axes Principle components
Principle components Easier to see right/left and above/below variation New X and Y axes
Principle components PC1 Easier to see right/left and above/below variation PC2 These new axes are called Principle Components or PCs New X and Y axes
Principle components PC1 Easier to see right/left and above/below variation PC1 spans along the direction of the most PC2 These new axes are called Principle Components or PCs New X and Y axes
Principle components PC1 Easier to see right/left and above/below variation PC1 spans along the direction of the most PC2 PC2 spans along the direction of the most These new axes are called Principle Components or PCs New X and Y axes
Principle components PC1 Easier to see right/left and above/below variation PC1 spans along the direction of the most PC2 PC2 spans along the direction of the most These new axes are called Principle Components or PCs Principle Components are always orthogonal one to another New X and Y axes
Principle components PC1 PC2 If original data would have 3 dimensions, we would have 3rd PC PC3
Principle components PC1 PC2 If original data would have 3 dimensions, we would have 3rd PC PC3 There is always as many PCs as there are dimensions
Principle components If original data would have 3 dimensions, we would have 3rd PC There is always as many PCs as there are dimensions Each new PCs is guaranteed to explain less variance PC1 PC2 PC3
Principle components If original data would have 3 dimensions, we would have 3rd PC There is always as many PCs as there are dimensions Usually it is enough to project data points onto first two PCs to see important patterns PC1 PC2 PC3 Each new PCs is guaranteed to explain less variance
PCA Now we know what those PC1 and PC2 stand for!
PCA Now we know what those PC1 and PC2 stand for! Wait! But so far we have been talking only about 2 proteins!
PCA Now we know what those PC1 and PC2 stand for! We are coming to this part Wait! But so far we have been talking only about 2 proteins!
Where PCs come from PC1 PC2
PC1 PC2 These are vectors Where PCs come from
PC1 PC2 30 3 Where PCs come from These are vectors All vectors have coordinates
PC1 PC2 30 3 For example, this one has coordinates (3,3) Where PCs come from These are vectors All vectors have coordinates
PC1 PC2 30 3 For example, this one has coordinates (3,3) In 200-D space vectors have 200 coordinates Where PCs come from These are vectors All vectors have coordinates
30 60 Protein#2 300 60 Protein #1 0 How did we choose PC1? Where PCs come from
30 60 Protein#2 300 60 Protein #1 0 How did we choose PC1? By minimising the distance from points to the vector Where PCs come from
30 60 Protein#2 300 60 Protein #1 0 How did we choose PC1? By minimising the distance from points to the vector Where PCs come from
30 60 Protein#2 300 60 Protein #1 0 How did we choose PC1? Coordinates of this vector are coordinates of PC1 By minimising the distance from points to the vector Where PCs come from PC1
30 60 Protein#2 300 60 Protein #1 0 How did we choose PC1? Coordinates of this vector are coordinates of PC1 By minimising the distance from points to the vector Where PCs come from Coordinates of PC2 are chosen in similar fashion but PC2 should be orthogonal to PC1 PC1 PC2
PCA Example Reactivities PCs coordinates PC1 PC2 2 3 4 12 … … -5 0 Let’s find coordinates of Person A in terms of PCs Person Pr1 Pr2 … PrN A 24 29 … 11 B 63 59 … 1 C 51 32 … 23 D 34 56 … 2 E 15 4 … 8 … … … … …
PCA Example Reactivities PCs coordinates PC1 PC2 2 3 4 12 … … -5 0 Let’s find coordinates of Person A in terms of PCs Person Pr1 Pr2 … PrN A 24 29 … 11 B 63 59 … 1 C 51 32 … 23 D 34 56 … 2 E 15 4 … 8 … … … … … asmanyasthere areproteins(N)
PCA Example Reactivities PCs coordinates Let’s find coordinates of Person A in terms of PCs Person A (PC1 score) = PC1 PC2 2 3 4 12 … … -5 0 Person Pr1 Pr2 … PrN A 24 29 … 11 B 63 59 … 1 C 51 32 … 23 D 34 56 … 2 E 15 4 … 8 … … … … …
PCA Example Reactivities PCs coordinates Let’s find coordinates of Person A in terms of PCs Person A (PC1 score) = 24*2 + 29*4 +… + 11*(-5) = 11 PC1 PC2 2 3 4 12 … … -5 0 Person Pr1 Pr2 … PrN A 24 29 … 11 B 63 59 … 1 C 51 32 … 23 D 34 56 … 2 E 15 4 … 8 … … … … …
PCA Example Reactivities PCs coordinates Person A (PC2 score) = 24*3+ 29*12 + … +11*0 = 21 Reactivities PCs coordinates Let’s find coordinates of Person A in terms of PCs Person A (PC1 score) = 24*2 + 29*4 +… + 11*(-5) = 11 PC1 PC2 2 3 4 12 … … -5 0 Person Pr1 Pr2 … PrN A 24 29 … 11 B 63 59 … 1 C 51 32 … 23 D 34 56 … 2 E 15 4 … 8 … … … … …
Principle components Principle component 1 Principle component 2 21 110 0 A Person A (PC2 score) = 24*3+ 29*12 + … +11*0 = 21 Person A (PC1 score) = 24*2 + 29*4 +… + 11*(-5) = 11
Principle components Principle component 1 Principle component 2 Person B (PC2 score) = 65 Person B (PC1 score) = 60 65 600 0 A B
Principle components Principle component 1 Principle component 2 0 0 A B C D E G
Principle components Principle component 1 Principle component 2 0 0 A B C D E G The idea is that similar points in multi-dimensional space get located close in fewer dimensions Healthy Patients
PCA is good, but it is a linear algorithm, meaning that it cannot capture complex relationship between features
PCA is good, but it is a linear algorithm, meaning that it cannot capture complex relationship between features There are always alternative options to consider…
https://www.analyticsvidhya.com/blog/2017/01/t-sne-implementation-r-python/ t-Distributed Stochastic Neighbour Embedding (t-SNE) visualisation is form http://distill.pub/2016/misread-tsne/ t-SNE is non-linear dimensionality reduction algorithm t-SNE uses conditional probabilities to represent similarity between points Has been shown to better represent underlying structure of multidimensional data
Practice
https://www.analyticsvidhya.com/blog/2017/01/t-sne-implementation-r-python/ t-SNE visualisation is form http://distill.pub/2016/misread-tsne/ PCA O(N2) makes it very slow Meaning of features is lost Has been shown to capture the structure of multi-dimensional data better (than PCA) Works relatively fast even on big datasets Transformed features could be traced back Usually is not so good in figuring out hidden structure
Demo
References • Machine Learning by Andrew Ng (https://www.coursera.org/learn/machine- learning) • Introduction to Machine Learning by Pascal Vincent given at Deep Learning Summer School, Montreal 2015 (http://videolectures.net/ deeplearning2015_vincent_machine_learning/) • Welcome to Machine Learning by Konstantin Tretyakov delivered at AACIMP Summer School 2015 (http://kt.era.ee/lectures/aacimp2015/1-intro.pdf) • Stanford CS class: Convolutional Neural Networks for Visual Recognition by Andrej Karpathy (http://cs231n.github.io/) • Data Mining Course by Jaak Vilo at University of Tartu (https://courses.cs.ut.ee/ MTAT.03.183/2017_spring/uploads/Main/DM_05_Clustering.pdf) • Machine Learning Essential Conepts by Ilya Kuzovkin (https:// www.slideshare.net/iljakuzovkin) • From the brain to deep learning and back by Raul Vicente Zafra and Ilya Kuzovkin (http://www.uttv.ee/naita?id=23585&keel=eng)
www.biit.cs.ut.ee www.ut.ee www.quretec.ee

More Related Content

PDF
CNN
georgejustymirobi1
 
PDF
High Dimensional Data Visualization using t-SNE
Kai-Wen Zhao
 
PDF
Optimization for Neural Network Training - Veronica Vilaplana - UPC Barcelona...
Universitat Politècnica de Catalunya
 
PDF
Visualizing Data Using t-SNE
David Khosid
 
PPTX
Hyperparameter Tuning
Jon Lederman
 
PDF
Dimensionality reduction with UMAP
Jakub Bartczuk
 
PPTX
Linear regression
zekeLabs Technologies
 
PDF
Data preprocessing using Machine Learning
Gopal Sakarkar
 
CNN
georgejustymirobi1
 
High Dimensional Data Visualization using t-SNE
Kai-Wen Zhao
 
Optimization for Neural Network Training - Veronica Vilaplana - UPC Barcelona...
Universitat Politècnica de Catalunya
 
Visualizing Data Using t-SNE
David Khosid
 
Hyperparameter Tuning
Jon Lederman
 
Dimensionality reduction with UMAP
Jakub Bartczuk
 
Linear regression
zekeLabs Technologies
 
Data preprocessing using Machine Learning
Gopal Sakarkar
 

What's hot (20)

PPTX
Naive Bayes Classifier | Naive Bayes Algorithm | Naive Bayes Classifier With ...
Simplilearn
 
PPT
Support Vector Machines
nextlib
 
PPTX
Deep Learning Frameworks 2019 | Which Deep Learning Framework To Use | Deep L...
Simplilearn
 
PDF
Beginners Guide to Non-Negative Matrix Factorization
Benjamin Bengfort
 
PDF
PR-315: Taming Transformers for High-Resolution Image Synthesis
Hyeongmin Lee
 
PPTX
ML - Multiple Linear Regression
Andrew Ferlitsch
 
PPTX
Data preprocessing in Machine learning
pyingkodi maran
 
PDF
Feature Engineering in Machine Learning
Knoldus Inc.
 
PPTX
Fuzzy logic and application in AI
Ildar Nurgaliev
 
PDF
Notes from Coursera Deep Learning courses by Andrew Ng
dataHacker. rs
 
PPTX
Particle filter
Mohammad Reza Jabbari
 
PPTX
Classification Algorithm.
Megha Sharma
 
PPTX
Gradient descent method
Prof. Neeta Awasthy
 
PPTX
support vector regression
Akhilesh Joshi
 
PDF
Nonnegative Matrix Factorization
Tatsuya Yokota
 
PPTX
Decision Tree Learning
Md. Ariful Hoque
 
PDF
Faster R-CNN: Towards real-time object detection with region proposal network...
Universitat Politècnica de Catalunya
 
PDF
Deep Learning: Application & Opportunity
iTrain
 
PPTX
Reinforcement learning
Ding Li
 
PDF
Introduction to Recurrent Neural Network
Yan Xu
 
Naive Bayes Classifier | Naive Bayes Algorithm | Naive Bayes Classifier With ...
Simplilearn
 
Support Vector Machines
nextlib
 
Deep Learning Frameworks 2019 | Which Deep Learning Framework To Use | Deep L...
Simplilearn
 
Beginners Guide to Non-Negative Matrix Factorization
Benjamin Bengfort
 
PR-315: Taming Transformers for High-Resolution Image Synthesis
Hyeongmin Lee
 
ML - Multiple Linear Regression
Andrew Ferlitsch
 
Data preprocessing in Machine learning
pyingkodi maran
 
Feature Engineering in Machine Learning
Knoldus Inc.
 
Fuzzy logic and application in AI
Ildar Nurgaliev
 
Notes from Coursera Deep Learning courses by Andrew Ng
dataHacker. rs
 
Particle filter
Mohammad Reza Jabbari
 
Classification Algorithm.
Megha Sharma
 
Gradient descent method
Prof. Neeta Awasthy
 
support vector regression
Akhilesh Joshi
 
Nonnegative Matrix Factorization
Tatsuya Yokota
 
Decision Tree Learning
Md. Ariful Hoque
 
Faster R-CNN: Towards real-time object detection with region proposal network...
Universitat Politècnica de Catalunya
 
Deep Learning: Application & Opportunity
iTrain
 
Reinforcement learning
Ding Li
 
Introduction to Recurrent Neural Network
Yan Xu
 
Ad

Similar to 4 Dimensionality reduction (PCA & t-SNE) (20)

PDF
26 soril dun_kodtoi
NARANGERELBAYASGALAN
 
PDF
26 soril dun_kodtoi_2
NARANGERELBAYASGALAN
 
PDF
41 soril dun.kodoor
NARANGERELBAYASGALAN
 
PDF
Ih soriliin dun 2013.05.25
Baasandorj Baatarhuyag
 
PDF
Informe simulacion digital yolfred uzcategui - 25.242.800
Yolfred Uzcategui
 
PDF
Ih sungaa 2013 dungiin jagsaalt
Baasandorj Baatarhuyag
 
PDF
40 soril dun_kod
NARANGERELBAYASGALAN
 
PPSX
Data Encryption Standard (DES)
Amir Masinaei
 
KEY
IP Addressing and Subnetting Basics
Rowell Dionicio
 
PPTX
Binary numbers
dmendonsa
 
PDF
Ip Addressing And Subnetting Teachers Book Robb Jones
Camposol S.A.
 
PDF
Ip addressing and subnetting instructors workbook
Swapnil Kapate
 
PDF
Ip Subredes
Espol
 
PDF
Soril 29 dun
NARANGERELBAYASGALAN
 
PDF
25 soril dun
NARANGERELBAYASGALAN
 
DOCX
Norma 2215105030(tgas 3)
Norma Mahmudah
 
PDF
Day02
avinashbolla
 
TXT
Chuongtrinh led ma trix
huy hung
 
PDF
Bits, Bytes and Blobs
Mrinal Wadhwa
 
DOCX
Conversiones binario decimal/ decimal binario
aidalyb2
 
26 soril dun_kodtoi
NARANGERELBAYASGALAN
 
26 soril dun_kodtoi_2
NARANGERELBAYASGALAN
 
41 soril dun.kodoor
NARANGERELBAYASGALAN
 
Ih soriliin dun 2013.05.25
Baasandorj Baatarhuyag
 
Informe simulacion digital yolfred uzcategui - 25.242.800
Yolfred Uzcategui
 
Ih sungaa 2013 dungiin jagsaalt
Baasandorj Baatarhuyag
 
40 soril dun_kod
NARANGERELBAYASGALAN
 
Data Encryption Standard (DES)
Amir Masinaei
 
IP Addressing and Subnetting Basics
Rowell Dionicio
 
Binary numbers
dmendonsa
 
Ip Addressing And Subnetting Teachers Book Robb Jones
Camposol S.A.
 
Ip addressing and subnetting instructors workbook
Swapnil Kapate
 
Ip Subredes
Espol
 
Soril 29 dun
NARANGERELBAYASGALAN
 
25 soril dun
NARANGERELBAYASGALAN
 
Norma 2215105030(tgas 3)
Norma Mahmudah
 
Day02
avinashbolla
 
Chuongtrinh led ma trix
huy hung
 
Bits, Bytes and Blobs
Mrinal Wadhwa
 
Conversiones binario decimal/ decimal binario
aidalyb2
 
Ad

More from Dmytro Fishman (14)

PDF
DOME: Recommendations for supervised machine learning validation in biology
Dmytro Fishman
 
PDF
Tips for effective presentations
Dmytro Fishman
 
PDF
Autonomous Driving Lab - Simultaneous Localization and Mapping WP
Dmytro Fishman
 
PDF
Introduction to Deep Learning
Dmytro Fishman
 
PDF
Introduction to Machine Learning for Taxify/Bolt
Dmytro Fishman
 
PDF
Introduction to Gaussian Processes
Dmytro Fishman
 
PDF
Biit group 2018
Dmytro Fishman
 
PDF
Detecting Nuclei from Microscopy Images with Deep Learning
Dmytro Fishman
 
PDF
Deep Learning in Healthcare
Dmytro Fishman
 
PDF
5 Introduction to neural networks
Dmytro Fishman
 
PDF
3 Unsupervised learning
Dmytro Fishman
 
PDF
1 Supervised learning
Dmytro Fishman
 
PDF
What does it mean to be a bioinformatician?
Dmytro Fishman
 
PDF
Machine Learning in Bioinformatics
Dmytro Fishman
 
DOME: Recommendations for supervised machine learning validation in biology
Dmytro Fishman
 
Tips for effective presentations
Dmytro Fishman
 
Autonomous Driving Lab - Simultaneous Localization and Mapping WP
Dmytro Fishman
 
Introduction to Deep Learning
Dmytro Fishman
 
Introduction to Machine Learning for Taxify/Bolt
Dmytro Fishman
 
Introduction to Gaussian Processes
Dmytro Fishman
 
Biit group 2018
Dmytro Fishman
 
Detecting Nuclei from Microscopy Images with Deep Learning
Dmytro Fishman
 
Deep Learning in Healthcare
Dmytro Fishman
 
5 Introduction to neural networks
Dmytro Fishman
 
3 Unsupervised learning
Dmytro Fishman
 
1 Supervised learning
Dmytro Fishman
 
What does it mean to be a bioinformatician?
Dmytro Fishman
 
Machine Learning in Bioinformatics
Dmytro Fishman
 

Recently uploaded (20)

PDF
Vision Prelims GS PYQ Analysis 2011-2022 www.upscpdf.com.pdf
navneetgupta711851
 
PDF
English Textual Question & Ans (12th Class).pdf
javaidpaswal33
 
PDF
LIFE & LIVING TRILOGY- PART (1) WHO ARE WE.pdf
sureshkumarsoni26
 
PDF
Myanmar Dental Journal, The Journal of the Myanmar Dental Association (2013).pdf
Nay Aung
 
PDF
LIFE & LIVING TRILOGY - PART - (2) THE PURPOSE OF LIFE.pdf
sureshkumarsoni26
 
PPTX
A powerpoint presentation on the Revised K-10 Science Shaping Paper
JericEstaco2
 
PDF
Climate and Adaptation MCQs class 7 from chatgpt
AkashDTejwani
 
PDF
Race Reva University – Shaping Future Leaders in Artificial Intelligence
RACE REVA University
 
PDF
My India Quiz Book_20210205121199924.pdf
AtandraDey2
 
PDF
HVAC Specification 2024 according to central public works department
amitrobanipl
 
PDF
MBA _Common_ 2nd year Syllabus _2021-22_.pdf
DrAnubha4
 
PDF
Journal of Dental Science - UDMY (2021).pdf
Nay Aung
 
PDF
BP 505 T. PHARMACEUTICAL JURISPRUDENCE (UNIT 1).pdf
CMEmpire
 
PDF
BP 704 T. NOVEL DRUG DELIVERY SYSTEMS (UNIT 2).pdf
CMEmpire
 
PDF
BP 505 T. PHARMACEUTICAL JURISPRUDENCE (UNIT 2).pdf
CMEmpire
 
PPTX
Climate Change and Its Global Impact.pptx
vaishalidobhal148
 
PPTX
Core Concepts of Personalized Learning and Virtual Learning Environments
Dr. Bushra Sumaiya
 
PDF
FOISHS ANNUAL IMPLEMENTATION PLAN 2025.pdf
EllenJoyCaniya2
 
PDF
Environmental Education MCQ BD2EE - Share Source.pdf
Dr Raja Mohammed T
 
PDF
FORM 1 BIOLOGY MIND MAPS and their schemes
arcade5
 
Vision Prelims GS PYQ Analysis 2011-2022 www.upscpdf.com.pdf
navneetgupta711851
 
English Textual Question & Ans (12th Class).pdf
javaidpaswal33
 
LIFE & LIVING TRILOGY- PART (1) WHO ARE WE.pdf
sureshkumarsoni26
 
Myanmar Dental Journal, The Journal of the Myanmar Dental Association (2013).pdf
Nay Aung
 
LIFE & LIVING TRILOGY - PART - (2) THE PURPOSE OF LIFE.pdf
sureshkumarsoni26
 
A powerpoint presentation on the Revised K-10 Science Shaping Paper
JericEstaco2
 
Climate and Adaptation MCQs class 7 from chatgpt
AkashDTejwani
 
Race Reva University – Shaping Future Leaders in Artificial Intelligence
RACE REVA University
 
My India Quiz Book_20210205121199924.pdf
AtandraDey2
 
HVAC Specification 2024 according to central public works department
amitrobanipl
 
MBA _Common_ 2nd year Syllabus _2021-22_.pdf
DrAnubha4
 
Journal of Dental Science - UDMY (2021).pdf
Nay Aung
 
BP 505 T. PHARMACEUTICAL JURISPRUDENCE (UNIT 1).pdf
CMEmpire
 
BP 704 T. NOVEL DRUG DELIVERY SYSTEMS (UNIT 2).pdf
CMEmpire
 
BP 505 T. PHARMACEUTICAL JURISPRUDENCE (UNIT 2).pdf
CMEmpire
 
Climate Change and Its Global Impact.pptx
vaishalidobhal148
 
Core Concepts of Personalized Learning and Virtual Learning Environments
Dr. Bushra Sumaiya
 
FOISHS ANNUAL IMPLEMENTATION PLAN 2025.pdf
EllenJoyCaniya2
 
Environmental Education MCQ BD2EE - Share Source.pdf
Dr Raja Mohammed T
 
FORM 1 BIOLOGY MIND MAPS and their schemes
arcade5
 

4 Dimensionality reduction (PCA & t-SNE)

  • 1. Introduction to Machine Learning (Dimensionality reduction) Dmytro Fishman (dmytro@ut.ee)
  • 2. So far we have been fearless drawing multi-dimensional instances on two-dimensional planes
  • 4. 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 126, 213, 34, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 19, 227, 254, 84, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 114, 254, 254, 84, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 178, 254, 254, 84, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 60, 236, 254, 254, 84, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 78, 254, 254, 254, 84, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 78, 254, 254, 231, 56, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 191, 254, 254, 185, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 233, 254, 254, 105, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 233, 254, 254, 30, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 93, 248, 254, 243, 27, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 16, 192, 254, 254, 130, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 32, 254, 254, 248, 95, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 41, 254, 254, 231, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 186, 254, 254, 160, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 21, 203, 254, 188, 12, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 86, 254, 255, 179, 8, 4, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 86, 254, 254, 254, 254, 130, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 68, 240, 254, 254, 250, 103, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 92, 255, 207, 91, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0 28px 28px
  • 5. 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 126, 213, 34, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 19, 227, 254, 84, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 114, 254, 254, 84, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 178, 254, 254, 84, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 60, 236, 254, 254, 84, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 78, 254, 254, 254, 84, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 78, 254, 254, 231, 56, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 191, 254, 254, 185, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 233, 254, 254, 105, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 233, 254, 254, 30, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 93, 248, 254, 243, 27, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 16, 192, 254, 254, 130, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 32, 254, 254, 248, 95, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 41, 254, 254, 231, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 186, 254, 254, 160, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 21, 203, 254, 188, 12, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 86, 254, 255, 179, 8, 4, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 86, 254, 254, 254, 254, 130, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 68, 240, 254, 254, 250, 103, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 92, 255, 207, 91, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0 28px 28px 102 41
  • 6. 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 126, 213, 34, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 19, 227, 254, 84, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 114, 254, 254, 84, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 178, 254, 254, 84, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 60, 236, 254, 254, 84, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 78, 254, 254, 254, 84, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 78, 254, 254, 231, 56, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 191, 254, 254, 185, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 233, 254, 254, 105, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 233, 254, 254, 30, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 93, 248, 254, 243, 27, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 16, 192, 254, 254, 130, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 32, 254, 254, 248, 95, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 41, 254, 254, 231, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 186, 254, 254, 160, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 21, 203, 254, 188, 12, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 86, 254, 255, 179, 8, 4, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 86, 254, 254, 254, 254, 130, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 68, 240, 254, 254, 250, 103, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 92, 255, 207, 91, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0 28px 28px 41 Pixel #X Pixel#Y 102
  • 7. 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 126, 213, 34, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 19, 227, 254, 84, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 114, 254, 254, 84, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 178, 254, 254, 84, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 60, 236, 254, 254, 84, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 78, 254, 254, 254, 84, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 78, 254, 254, 231, 56, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 191, 254, 254, 185, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 233, 254, 254, 105, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 233, 254, 254, 30, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 93, 248, 254, 243, 27, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 16, 192, 254, 254, 130, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 32, 254, 254, 248, 95, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 41, 254, 254, 231, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 186, 254, 254, 160, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 21, 203, 254, 188, 12, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 86, 254, 255, 179, 8, 4, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 86, 254, 254, 254, 254, 130, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 68, 240, 254, 254, 250, 103, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 92, 255, 207, 91, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0 28px 28px Pixel #X Pixel#Y 41 How about the rest of 782 pixels? 102
  • 8. 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 126, 213, 34, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 19, 227, 254, 84, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 114, 254, 254, 84, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 178, 254, 254, 84, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 60, 236, 254, 254, 84, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 78, 254, 254, 254, 84, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 78, 254, 254, 231, 56, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 191, 254, 254, 185, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 233, 254, 254, 105, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 233, 254, 254, 30, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 93, 248, 254, 243, 27, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 16, 192, 254, 254, 130, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 32, 254, 254, 248, 95, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 41, 254, 254, 231, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 186, 254, 254, 160, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 21, 203, 254, 188, 12, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 86, 254, 255, 179, 8, 4, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 86, 254, 254, 254, 254, 130, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 68, 240, 254, 254, 250, 103, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 92, 255, 207, 91, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0 28px 28px Pixel #X Pixel#Y 41 How about the rest of 782 pixels? But we cannot really visualise all 784 dimensions... 102
  • 9. 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 126, 213, 34, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 19, 227, 254, 84, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 114, 254, 254, 84, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 178, 254, 254, 84, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 60, 236, 254, 254, 84, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 78, 254, 254, 254, 84, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 78, 254, 254, 231, 56, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 191, 254, 254, 185, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 233, 254, 254, 105, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 233, 254, 254, 30, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 93, 248, 254, 243, 27, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 16, 192, 254, 254, 130, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 32, 254, 254, 248, 95, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 41, 254, 254, 231, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 186, 254, 254, 160, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 21, 203, 254, 188, 12, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 86, 254, 255, 179, 8, 4, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 86, 254, 254, 254, 254, 130, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 68, 240, 254, 254, 250, 103, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 92, 255, 207, 91, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0 28px 28px Pixel #X Pixel#Y 41 How about the rest of 782 pixels? 102 But we cannot really visualise all 784 dimensions... or can we?
  • 11. The following material was inspired by: https://www.youtube.com/watch?v=_UVHneBUBW0
  • 12. The following material was inspired by: https://www.youtube.com/watch?v=_UVHneBUBW0 With a bit of fixes, though
  • 14. Patients Healthy Each dot in this plot represents a person
  • 15. Patients Healthy Every person in its turn could be described by hundreds of protein reactivities Each dot in this plot represents a person
  • 16. Patients Healthy Every person in its turn could be described by hundreds of protein reactivities Each dot in this plot represents a person The idea is that people with similar protein reactivity profiles cluster together
  • 17. Patients Healthy 5, 98, 43, …, 9, 66, Patient
  • 18. Patients Healthy 5, 98, 43, …, 9, 66, Patient 62, 8, 14, …, 73, 2, Healthy
  • 19. Patients Healthy 5, 98, 43, …, 9, 66, Patient 62, 8, 14, …, 73, 2, Healthy 3, 45, 13, …, 7, 54, Patient
  • 20. Patients Healthy 5, 98, 43, …, 9, 66, Patient 62, 8, 14, …, 73, 2, Healthy 3, 45, 13, …, 7, 54, Patient How is that even possible?
  • 21. Patients Healthy PCA 5, 98, 43, …, 9, 66, Patient 62, 8, 14, …, 73, 2, Healthy 3, 45, 13, …, 7, 54, Patient How is that even possible?
  • 23. 1-Dimensional data Person Protein #1 A 24 B 63 C 51 D 34 E 15
  • 24. 1-Dimensional data 300 60 Person Protein #1 A 24 B 63 C 51 D 34 E 15
  • 25. 1-Dimensional data 300 60 300 60 Uniformly distributed data Person Protein #1 A 24 B 63 C 51 D 34 E 15
  • 26. 1-Dimensional data 300 60 300 60 Uniformly distributed data 30 60 Non-uniform distribution 0 Person Protein #1 A 24 B 63 C 51 D 34 E 15
  • 27. 2-Dimensional data 300 60 Protein #1 Person Protein #1 A 24 B 63 C 51 D 34 E 15
  • 28. 2-Dimensional data 30 60 Protein#2 Person Protein #1 Protein #2 A 24 B 63 C 51 D 34 E 15 300 60 Protein #1 0
  • 29. 2-Dimensional data Person Protein #1 Protein #2 A 24 29 B 63 C 51 D 34 E 15 A30 60 Protein#2 300 60 Protein #1 0
  • 30. 2-Dimensional data A Person Protein #1 Protein #2 A 24 29 B 63 C 51 D 34 E 15 30 60 Protein#2 300 60 Protein #1 0
  • 31. 2-Dimensional data A Person Protein #1 Protein #2 A 24 29 B 63 59 C 51 D 34 E 15 B 30 60 Protein#2 300 60 Protein #1 0
  • 32. 2-Dimensional data A B And so on… Person Protein #1 Protein #2 A 24 29 B 63 59 C 51 D 34 E 15 30 60 Protein#2 300 60 Protein #1 0
  • 33. 2-Dimensional data Person Protein #1 Protein #2 A 24 29 B 63 59 C 51 32 D 34 56 E 15 4 … … … 30 60 Protein#2 300 60 Protein #1 0
  • 34. 2-Dimensional data Protein profiles are correlated Person Protein #1 Protein #2 A 24 29 B 63 59 C 51 32 D 34 56 E 15 4 … … … 30 60 Protein#2 300 60 Protein #1 0
  • 35. 2-Dimensional data Here is no correlation Person Protein #1 Protein #2 A 24 29 B 63 59 C 51 32 D 34 56 E 15 4 … … … 30 60 Protein#2 300 60 Protein #1 0
  • 36. 3-Dimensional data Person Protein #1 Protein #2 A 24 29 B 63 59 C 51 32 D 34 56 E 15 4 … … … 30 60 Protein#2 300 60 Protein #1 0
  • 37. 3-Dimensional data Person P1 P2 P3 A 24 29 B 63 59 C 51 32 D 34 56 E 15 4 … … … Protein #330 60 Protein#2 300 60 Protein #1 0 30 60
  • 38. 3-Dimensional data Person P1 P2 P3 A 24 29 33 B 63 59 C 51 32 D 34 56 E 15 4 … … … A 30 60 Protein #330 60 Protein#2 300 60 Protein #1 0
  • 41. Dimensions 300 60 30 6 3 6 3-Dimensional 30 6 3 6 3 6 2-Dimensional 1-Dimensional
  • 43. Are all D are equally important? How about 200-D data?
  • 44. Are all D are equally important? Let’s go back to 2-D example How about 200-D data?
  • 46. Variation is from left to right 30 60 Protein#2 300 60 Protein #1 0 Importance of axes
  • 47. Projected data - does not look too different 30 60 Protein#2 300 60 Protein #1 0 Importance of axes
  • 48. We can safely remove 2nd dimension 30 60 Protein#2 300 60 Protein #1 0 300 60 Protein #1 Importance of axes Projected data - does not look too different
  • 49. 30 60 Protein#2 300 60 Protein #1 0 300 60 Protein #1 The important variation is from left to right We can safely remove 2nd dimension Importance of axes Projected data - does not look too different
  • 51. 30 60 Protein#2 300 60 Protein #1 0 Data is mostly spread along this line Principle components
  • 52. 30 60 Protein#2 300 60 Protein #1 0 Data is mostly spread along this line and a little bit along this line Principle components
  • 53. 30 60 Protein#2 300 60 Protein #1 0 Data is mostly spread along this line and a little bit along this line What if we make new axes from these lines? Principle components
  • 54. 30 60 Protein #2 30 0 60 Protein #1 0 New X and Y axes Principle components
  • 55. Principle components Easier to see right/left and above/below variation New X and Y axes
  • 56. Principle components PC1 Easier to see right/left and above/below variation PC2 These new axes are called Principle Components or PCs New X and Y axes
  • 57. Principle components PC1 Easier to see right/left and above/below variation PC1 spans along the direction of the most PC2 These new axes are called Principle Components or PCs New X and Y axes
  • 58. Principle components PC1 Easier to see right/left and above/below variation PC1 spans along the direction of the most PC2 PC2 spans along the direction of the most These new axes are called Principle Components or PCs New X and Y axes
  • 59. Principle components PC1 Easier to see right/left and above/below variation PC1 spans along the direction of the most PC2 PC2 spans along the direction of the most These new axes are called Principle Components or PCs Principle Components are always orthogonal one to another New X and Y axes
  • 60. Principle components PC1 PC2 If original data would have 3 dimensions, we would have 3rd PC PC3
  • 61. Principle components PC1 PC2 If original data would have 3 dimensions, we would have 3rd PC PC3 There is always as many PCs as there are dimensions
  • 62. Principle components If original data would have 3 dimensions, we would have 3rd PC There is always as many PCs as there are dimensions Each new PCs is guaranteed to explain less variance PC1 PC2 PC3
  • 63. Principle components If original data would have 3 dimensions, we would have 3rd PC There is always as many PCs as there are dimensions Usually it is enough to project data points onto first two PCs to see important patterns PC1 PC2 PC3 Each new PCs is guaranteed to explain less variance
  • 64. PCA Now we know what those PC1 and PC2 stand for!
  • 65. PCA Now we know what those PC1 and PC2 stand for! Wait! But so far we have been talking only about 2 proteins!
  • 66. PCA Now we know what those PC1 and PC2 stand for! We are coming to this part Wait! But so far we have been talking only about 2 proteins!
  • 67. Where PCs come from PC1 PC2
  • 69. PC1 PC2 30 3 Where PCs come from These are vectors All vectors have coordinates
  • 70. PC1 PC2 30 3 For example, this one has coordinates (3,3) Where PCs come from These are vectors All vectors have coordinates
  • 71. PC1 PC2 30 3 For example, this one has coordinates (3,3) In 200-D space vectors have 200 coordinates Where PCs come from These are vectors All vectors have coordinates
  • 72. 30 60 Protein#2 300 60 Protein #1 0 How did we choose PC1? Where PCs come from
  • 73. 30 60 Protein#2 300 60 Protein #1 0 How did we choose PC1? By minimising the distance from points to the vector Where PCs come from
  • 74. 30 60 Protein#2 300 60 Protein #1 0 How did we choose PC1? By minimising the distance from points to the vector Where PCs come from
  • 75. 30 60 Protein#2 300 60 Protein #1 0 How did we choose PC1? Coordinates of this vector are coordinates of PC1 By minimising the distance from points to the vector Where PCs come from PC1
  • 76. 30 60 Protein#2 300 60 Protein #1 0 How did we choose PC1? Coordinates of this vector are coordinates of PC1 By minimising the distance from points to the vector Where PCs come from Coordinates of PC2 are chosen in similar fashion but PC2 should be orthogonal to PC1 PC1 PC2
  • 77. PCA Example Reactivities PCs coordinates PC1 PC2 2 3 4 12 … … -5 0 Let’s find coordinates of Person A in terms of PCs Person Pr1 Pr2 … PrN A 24 29 … 11 B 63 59 … 1 C 51 32 … 23 D 34 56 … 2 E 15 4 … 8 … … … … …
  • 78. PCA Example Reactivities PCs coordinates PC1 PC2 2 3 4 12 … … -5 0 Let’s find coordinates of Person A in terms of PCs Person Pr1 Pr2 … PrN A 24 29 … 11 B 63 59 … 1 C 51 32 … 23 D 34 56 … 2 E 15 4 … 8 … … … … … asmanyasthere areproteins(N)
  • 79. PCA Example Reactivities PCs coordinates Let’s find coordinates of Person A in terms of PCs Person A (PC1 score) = PC1 PC2 2 3 4 12 … … -5 0 Person Pr1 Pr2 … PrN A 24 29 … 11 B 63 59 … 1 C 51 32 … 23 D 34 56 … 2 E 15 4 … 8 … … … … …
  • 80. PCA Example Reactivities PCs coordinates Let’s find coordinates of Person A in terms of PCs Person A (PC1 score) = 24*2 + 29*4 +… + 11*(-5) = 11 PC1 PC2 2 3 4 12 … … -5 0 Person Pr1 Pr2 … PrN A 24 29 … 11 B 63 59 … 1 C 51 32 … 23 D 34 56 … 2 E 15 4 … 8 … … … … …
  • 81. PCA Example Reactivities PCs coordinates Person A (PC2 score) = 24*3+ 29*12 + … +11*0 = 21 Reactivities PCs coordinates Let’s find coordinates of Person A in terms of PCs Person A (PC1 score) = 24*2 + 29*4 +… + 11*(-5) = 11 PC1 PC2 2 3 4 12 … … -5 0 Person Pr1 Pr2 … PrN A 24 29 … 11 B 63 59 … 1 C 51 32 … 23 D 34 56 … 2 E 15 4 … 8 … … … … …
  • 82. Principle components Principle component 1 Principle component 2 21 110 0 A Person A (PC2 score) = 24*3+ 29*12 + … +11*0 = 21 Person A (PC1 score) = 24*2 + 29*4 +… + 11*(-5) = 11
  • 83. Principle components Principle component 1 Principle component 2 Person B (PC2 score) = 65 Person B (PC1 score) = 60 65 600 0 A B
  • 84. Principle components Principle component 1 Principle component 2 0 0 A B C D E G
  • 85. Principle components Principle component 1 Principle component 2 0 0 A B C D E G The idea is that similar points in multi-dimensional space get located close in fewer dimensions Healthy Patients
  • 86. PCA is good, but it is a linear algorithm, meaning that it cannot capture complex relationship between features
  • 87. PCA is good, but it is a linear algorithm, meaning that it cannot capture complex relationship between features There are always alternative options to consider…
  • 88. https://www.analyticsvidhya.com/blog/2017/01/t-sne-implementation-r-python/ t-Distributed Stochastic Neighbour Embedding (t-SNE) visualisation is form http://distill.pub/2016/misread-tsne/ t-SNE is non-linear dimensionality reduction algorithm t-SNE uses conditional probabilities to represent similarity between points Has been shown to better represent underlying structure of multidimensional data
  • 90. https://www.analyticsvidhya.com/blog/2017/01/t-sne-implementation-r-python/ t-SNE visualisation is form http://distill.pub/2016/misread-tsne/ PCA O(N2) makes it very slow Meaning of features is lost Has been shown to capture the structure of multi-dimensional data better (than PCA) Works relatively fast even on big datasets Transformed features could be traced back Usually is not so good in figuring out hidden structure
  • 91. Demo
  • 92. References • Machine Learning by Andrew Ng (https://www.coursera.org/learn/machine- learning) • Introduction to Machine Learning by Pascal Vincent given at Deep Learning Summer School, Montreal 2015 (http://videolectures.net/ deeplearning2015_vincent_machine_learning/) • Welcome to Machine Learning by Konstantin Tretyakov delivered at AACIMP Summer School 2015 (http://kt.era.ee/lectures/aacimp2015/1-intro.pdf) • Stanford CS class: Convolutional Neural Networks for Visual Recognition by Andrej Karpathy (http://cs231n.github.io/) • Data Mining Course by Jaak Vilo at University of Tartu (https://courses.cs.ut.ee/ MTAT.03.183/2017_spring/uploads/Main/DM_05_Clustering.pdf) • Machine Learning Essential Conepts by Ilya Kuzovkin (https:// www.slideshare.net/iljakuzovkin) • From the brain to deep learning and back by Raul Vicente Zafra and Ilya Kuzovkin (http://www.uttv.ee/naita?id=23585&keel=eng)