1
00:00:07,902 --> 00:00:12,504
As you've probably gathered by following
the variety of topics in this course,

2
00:00:12,504 --> 00:00:16,902
the world of machine learning encompasses
a broad set of ideas and methods and

3
00:00:16,902 --> 00:00:19,691
the field itself continues
to advance rapidly.

4
00:00:19,691 --> 00:00:23,731
So the goal of a survey course like this
one is to cover the most important basic

5
00:00:23,731 --> 00:00:27,540
concepts and visit a few of the more
important, and interesting areas.

6
00:00:27,540 --> 00:00:28,506
And hopefully,

7
00:00:28,506 --> 00:00:32,652
inspire you enough to continue
exploring after you finish the course.

8
00:00:32,652 --> 00:00:33,982
I've put together this course.

9
00:00:33,982 --> 00:00:37,487
So it's not only a source of lectures and
assignments, but

10
00:00:37,487 --> 00:00:42,254
also a collection of useful readings you
can draw on based on my experience in this

11
00:00:42,254 --> 00:00:46,039
area that I've chose for
you as good starting points to continue

12
00:00:46,039 --> 00:00:49,502
your learning in different
aspects of machine learning.

13
00:00:49,502 --> 00:00:51,574
By now,
we've covered quite a bit of ground.

14
00:00:51,574 --> 00:00:55,447
Starting with an understanding of the
concepts and workflow of applied machine

15
00:00:55,447 --> 00:00:59,434
learning, and then getting some exposure
to a variety of learning algorithms for

16
00:00:59,434 --> 00:01:03,652
important tasks like classification, and
regression along with an understanding of

17
00:01:03,652 --> 00:01:07,779
the important parameters that control
modeling complexity for these algorithms.

18
00:01:09,010 --> 00:01:12,780
You should now have a better understanding
of problems like overfitting and

19
00:01:12,780 --> 00:01:17,250
data leakage along with an idea for some
strategies you can use for detecting, and

20
00:01:17,250 --> 00:01:18,290
avoiding these problems.

21
00:01:19,330 --> 00:01:23,060
We've covered how machine learning
algorithms ideas are evaluated and

22
00:01:23,060 --> 00:01:27,339
how to tune their parameters to optimize
different evaluation criteria that may be

23
00:01:27,339 --> 00:01:29,190
appropriate for different tasks.

24
00:01:29,190 --> 00:01:33,337
Finally, we've applied these ideas
with the help of notebook examples and

25
00:01:33,337 --> 00:01:38,004
assignments to gain more experience with a
very powerful machine learning library in

26
00:01:38,004 --> 00:01:39,521
Python, scikit-learn.

27
00:01:39,521 --> 00:01:43,279
I want to say something briefly about an
important family of machine learning tasks

28
00:01:43,279 --> 00:01:46,768
that we've touched on in this course,
but didn't focus on specifically and

29
00:01:46,768 --> 00:01:48,880
that is machine learning
that involves text.

30
00:01:49,950 --> 00:01:51,200
This includes problems,

31
00:01:51,200 --> 00:01:55,220
such as email spam detection and
classifying web pages by topic.

32
00:01:56,470 --> 00:01:59,610
The course following this one
in our data sign series focuses,

33
00:01:59,610 --> 00:02:03,300
specifically on text mining and
machine running with text.

34
00:02:03,300 --> 00:02:06,680
So if you have a particular interest in
diving into the use of natural language

35
00:02:06,680 --> 00:02:09,580
processing with machine learning,
I encourage you to check it out.

36
00:02:11,250 --> 00:02:13,220
Thanks for your interest in this course.

37
00:02:13,220 --> 00:02:17,050
I've really enjoyed putting together this
tour through the fascinating field of

38
00:02:17,050 --> 00:02:18,000
machine learning.

39
00:02:18,000 --> 00:02:21,913
I hope this course helped you in some way
along whatever path of exploration you

40
00:02:21,913 --> 00:02:22,573
might have.

41
00:02:22,573 --> 00:02:26,772
And as always, we welcome your
feedback on any aspect of this course.

42
00:02:26,772 --> 00:02:29,660
Thank you and
all the best in your own future learning.