Would you like to inspect the original subtitles? These are the user uploaded subtitles that are being translated:
1
00:00:01,260 --> 00:00:04,470
Machine learning projects can cover many different topics.
2
00:00:04,530 --> 00:00:09,930
It's important to design a framework you can use to approach different kinds of problems.
3
00:00:10,020 --> 00:00:15,390
You can consider what we're about to go through as like a little field guide that you can use for machine
4
00:00:15,390 --> 00:00:16,200
learning.
5
00:00:16,200 --> 00:00:20,470
So when you come up against a problem you can refer back to this field guide and go.
6
00:00:20,580 --> 00:00:21,070
Hold on.
7
00:00:21,090 --> 00:00:23,790
I need to break this problem down into a few little steps.
8
00:00:23,790 --> 00:00:30,390
What does a field guide say the framework we're going to be using comprises six steps.
9
00:00:30,390 --> 00:00:35,820
After working on many machine learning projects across multiple different industries these are the steps
10
00:00:35,880 --> 00:00:39,730
I found which come up time and time again.
11
00:00:39,900 --> 00:00:44,220
We're going to see this this diagram a lot in the next few lectures but in this one we're gonna We're
12
00:00:44,220 --> 00:00:50,750
gonna dive into each of these steps individually and and see what what kind of components they have.
13
00:00:50,770 --> 00:00:53,090
Step one is Problem Definition.
14
00:00:53,350 --> 00:00:59,230
Since we'll be focused on code first practical solutions it's important to define what problem we're
15
00:00:59,230 --> 00:01:00,520
trying to solve.
16
00:01:00,550 --> 00:01:03,400
Is it a supervised or unsupervised learning problem.
17
00:01:03,430 --> 00:01:06,540
Is it a classification or regression problem.
18
00:01:06,580 --> 00:01:11,730
Don't worry we'll see how to figure these out in the next few lectures.
19
00:01:11,750 --> 00:01:19,160
Step two is data since machine learning involves using algorithms to find and learn different patterns
20
00:01:19,160 --> 00:01:20,180
in data.
21
00:01:20,180 --> 00:01:24,480
Data is a requirement for any machine learning project.
22
00:01:24,550 --> 00:01:29,080
The question we're trying to answer here in step two is what kind of data do we have.
23
00:01:29,150 --> 00:01:36,110
Depending on the problem there are different kinds of data structure data such as rows and columns or
24
00:01:36,230 --> 00:01:43,410
what you'd expect to find in an Excel spreadsheet or unstructured data such as images or audio.
25
00:01:43,430 --> 00:01:50,300
Once we know what kind of data we have we can start to make decisions on how to use machine learning
26
00:01:50,300 --> 00:01:58,340
with it Step three is evaluation here will define what success means to us.
27
00:01:58,490 --> 00:02:04,580
Since machine learning since much of machine learning actually is experimental you could keep going
28
00:02:04,580 --> 00:02:08,810
forever trying to improve your results in search of the perfect model.
29
00:02:09,560 --> 00:02:15,530
However since we are practitioners we know the perfect model doesn't exist.
30
00:02:15,530 --> 00:02:22,430
Instead we begin by saying for this machine learning real estate project to be feasible we need at least
31
00:02:22,520 --> 00:02:27,620
a 95 percent accurate model at predicting the cost of houses.
32
00:02:27,620 --> 00:02:34,420
Of course in the beginning this evaluation metric won't be exact and will likely change over time.
33
00:02:34,460 --> 00:02:42,260
But having this at the start of a project gives us something to aim for Step 4 is features.
34
00:02:42,340 --> 00:02:46,410
The question we answer here is what do we already know about the data.
35
00:02:46,410 --> 00:02:51,360
Now even within different types of data there are different kinds of features.
36
00:02:51,490 --> 00:02:57,850
For example for predicting whether or not someone has heart disease you might use their body weight
37
00:02:57,880 --> 00:03:01,090
as a feature since body weight is a number.
38
00:03:01,390 --> 00:03:08,050
It's called a numerical feature and after talking to a doctor they might tell you if someone's body
39
00:03:08,050 --> 00:03:10,230
weight is over a certain number.
40
00:03:10,300 --> 00:03:14,030
They're more likely to have heart disease.
41
00:03:14,170 --> 00:03:18,550
There are more kinds of features such as categorical and derived.
42
00:03:18,550 --> 00:03:20,590
We're going to look at these in future lessons.
43
00:03:20,830 --> 00:03:28,030
But the premise remains a machine learning algorithms goal is to turn these features such as weight
44
00:03:28,240 --> 00:03:36,130
sex blood pressure and chest pain into patterns to make predictions such as whether or not a patient.
45
00:03:36,240 --> 00:03:43,770
We've got unique patient ideas here has heart disease or not Step five is modelling.
46
00:03:43,890 --> 00:03:48,670
Once you've learned a little bit about your data the next step is to model it.
47
00:03:48,980 --> 00:03:55,240
The question here is based on our problem and data what machine learning model should we use.
48
00:03:55,320 --> 00:03:59,970
Unlike other algorithms and sets of instructions you have to write from scratch.
49
00:04:00,000 --> 00:04:05,760
Many of the most useful machine learning algorithms have already been coded for you which is beautiful
50
00:04:05,760 --> 00:04:06,960
for us.
51
00:04:06,960 --> 00:04:12,900
Some models work better on different problems in others and in the beginning your focus will be to figure
52
00:04:12,900 --> 00:04:17,080
out the right model for the right kind of problem.
53
00:04:17,400 --> 00:04:19,660
Step six is experimentation.
54
00:04:19,890 --> 00:04:23,700
All of the steps we've just been through happen in a cycle.
55
00:04:23,700 --> 00:04:29,310
You might start out with one problem definition and find your data isn't suited to it then you might
56
00:04:29,310 --> 00:04:34,680
build a model and find it doesn't work as well as you outlined in your evaluation metric.
57
00:04:35,340 --> 00:04:40,690
So you build another one and you find out this one actually works pretty good.
58
00:04:40,730 --> 00:04:45,480
What's important to remember is although these steps are here those steps that we've been through in
59
00:04:45,480 --> 00:04:51,400
this framework it doesn't mean that they have to be followed in order nor are they set in stone.
60
00:04:51,420 --> 00:04:55,780
Consider them a rough guide now we've been through each of them briefly.
61
00:04:55,960 --> 00:04:58,350
Let's look at each one in a little bit more detail.
6793
Can't find what you're looking for?
Get subtitles in any language from opensubtitles.com, and translate them here.