1
00:00:01,060 --> 00:00:02,500
Welcome back.

2
00:00:02,500 --> 00:00:09,580
Up until now I've been blabbering on about what machine learning is and I hope you have a bit of an

3
00:00:09,580 --> 00:00:14,050
idea of why we have it and why it's useful.

4
00:00:14,170 --> 00:00:21,040
But at the end of the day the only reason that we care about machine learning is that we're able to

5
00:00:21,040 --> 00:00:26,290
use machines to predict results based on incoming data.

6
00:00:26,290 --> 00:00:27,310
That's it.

7
00:00:27,340 --> 00:00:29,440
Now this idea of machine learning.

8
00:00:29,440 --> 00:00:30,240
And don't worry.

9
00:00:30,280 --> 00:00:35,740
I know we keep talking about theory but I promise we're gonna get some coding exercises first.

10
00:00:35,740 --> 00:00:43,330
But I do want to talk quickly about some of the machine learning categories that you often see and keep

11
00:00:43,330 --> 00:00:51,420
in mind what I said that is machine learning is simply about predicting results based on incoming data.

12
00:00:51,640 --> 00:00:55,820
And all these subcategories simply do that.

13
00:00:55,840 --> 00:01:03,820
For example we have the idea of a supervised learning which is a subset of machine learning in this

14
00:01:03,880 --> 00:01:05,270
supervised learning.

15
00:01:05,500 --> 00:01:09,400
The data that we received already has CATEGORIES.

16
00:01:09,400 --> 00:01:13,740
THINK OF IT AS A CSC files with rows and columns label.

17
00:01:13,840 --> 00:01:22,410
We have labeled data and a test data that is label so we know if our function is right or wrong.

18
00:01:22,480 --> 00:01:30,310
So in a supervised learning scenario we can do things like classification to decide is this an apple.

19
00:01:30,310 --> 00:01:38,020
Or is this a paper machine learning model simply draws a line to decide Hey this is an apple and this

20
00:01:38,110 --> 00:01:46,360
is a pair or we might do something called regression instead of classification based on inputs.

21
00:01:46,550 --> 00:01:54,680
For example predicting stock prices another way that we might use supervised learning is for example

22
00:01:54,680 --> 00:02:03,740
to hire engineers based on inputs based on years of experience based on maybe age maybe where they live

23
00:02:04,040 --> 00:02:11,810
what type of computers they have all these sorts of inputs that are labeled can be used in a supervised

24
00:02:11,870 --> 00:02:17,570
learning system to decide should I hire this engineer or should I not.

25
00:02:17,600 --> 00:02:25,370
Now sometimes we have data that doesn't have labels and this is called on supervised learning.

26
00:02:25,370 --> 00:02:32,690
Again think of it as a CSP file without perhaps the column names labeled sometimes with things like

27
00:02:32,690 --> 00:02:33,820
clustering.

28
00:02:33,980 --> 00:02:39,170
We need to create these groups or at least the machine to create these groups.

29
00:02:39,200 --> 00:02:45,490
For example we give it a bunch of data points and then the machine decides oh this is a group.

30
00:02:45,530 --> 00:02:52,550
This is a group and this is a group or we can have something like association rule learning where we

31
00:02:52,550 --> 00:03:00,890
associate different things to predict what a customer perhaps might buy in the future when groups don't

32
00:03:00,890 --> 00:03:01,760
exist.

33
00:03:01,760 --> 00:03:04,200
We call it unsupervised learning.

34
00:03:04,310 --> 00:03:10,910
We can't tell the machine that they are right or wrong like we can when we do apples versus pears since

35
00:03:11,000 --> 00:03:17,050
there are no true categories but we let the machines just create these categories for us.

36
00:03:17,060 --> 00:03:24,010
Finally we have this idea of reinforcement learning and reinforcement learning is really interesting

37
00:03:24,040 --> 00:03:31,130
because it's all about teaching machines through trial and error through rewards and punishment so the

38
00:03:31,130 --> 00:03:38,930
program simply learns a game by playing it millions of times until well it gets the highest score it

39
00:03:38,930 --> 00:03:45,650
doesn't know what it's doing at first but then it tries to maximize the score and eventually figures

40
00:03:45,650 --> 00:03:51,270
out that hey maybe I should try and move where the ball is coming.

41
00:03:51,860 --> 00:03:58,250
So this is seen for skill acquisition or real time learning and you see it a lot in machine learning

42
00:03:58,250 --> 00:04:00,960
programs that play for example video games.

43
00:04:01,010 --> 00:04:08,210
But the idea here is that machine learning has different categories and different ways to accomplish

44
00:04:08,240 --> 00:04:17,830
its goal and topics like neural networks decision trees support vector machines K nearest neighbor are

45
00:04:17,830 --> 00:04:27,580
simply algorithms that are used with these sub fields in order to come to these predictions but remember

46
00:04:27,820 --> 00:04:35,650
the key thing all of these what they're doing is trying to learn from the data that it receives and

47
00:04:35,650 --> 00:04:40,570
predict something I'll see in the next one by.