1
00:00:00,000 --> 00:00:04,682
[MUSIC]

2
00:00:04,682 --> 00:00:08,270
Welcome to the first module
of the classification course.

3
00:00:08,270 --> 00:00:10,920
In this module we're going to
talk about linear classifiers,

4
00:00:10,920 --> 00:00:15,230
which are one of the most commonly
used classifiers out there.

5
00:00:15,230 --> 00:00:19,610
In particular, we're going to talk about
logistic regression, which is again one of

6
00:00:19,610 --> 00:00:24,310
the most commonly used linear classifiers
and one of the most useful ones.

7
00:00:24,310 --> 00:00:28,110
The concepts you learn here will actually
extend to lot of other classification

8
00:00:28,110 --> 00:00:30,900
methods beyond linear classifiers.

9
00:00:30,900 --> 00:00:33,850
You're going to learn
the fundamental concepts, but

10
00:00:33,850 --> 00:00:37,590
also the underlying algorithms

11
00:00:37,590 --> 00:00:41,280
that let you optimize the parameters of
this models to fit your training data.

12
00:00:42,550 --> 00:00:46,150
We'll start with motivating example,
the same one we used in the first

13
00:00:46,150 --> 00:00:50,440
course of the specialization,
a new kind of restaurant review system.

14
00:00:51,610 --> 00:00:54,330
Now there's a special day coming up for
me, and I want to go out for

15
00:00:54,330 --> 00:00:56,310
some great Japanese foods.

16
00:00:56,310 --> 00:01:00,600
And I want a really great restaurant for
it where I can eat amazing sushi.

17
00:01:00,600 --> 00:01:04,030
Seattle is a really great place
where we have a ton a different,

18
00:01:04,030 --> 00:01:07,180
highly rated Japanese restaurants.

19
00:01:07,180 --> 00:01:08,980
So how do I choose one?

20
00:01:08,980 --> 00:01:12,470
Well there are many aspects of restaurants
and many aspects that make them great.

21
00:01:12,470 --> 00:01:17,760
What I really care about is
amazing food and nice ambiance.

22
00:01:17,760 --> 00:01:19,170
If I look at a restaurant review,

23
00:01:19,170 --> 00:01:22,200
it will talk about different
aspects of the restaurant.

24
00:01:22,200 --> 00:01:26,220
So for example, a specific review might
have a sentence that says something like,

25
00:01:26,220 --> 00:01:28,270
the experience there was excellent.

26
00:01:28,270 --> 00:01:29,470
What does this tell me?

27
00:01:29,470 --> 00:01:31,210
Well if I look at this sentence,

28
00:01:31,210 --> 00:01:35,280
the sentence is positive about the
experience of going to this restaurant.

29
00:01:35,280 --> 00:01:38,148
Then there might be a next
sentence that says,

30
00:01:38,148 --> 00:01:41,542
my wife tried the ramen and
it was pretty forgettable.

31
00:01:41,542 --> 00:01:43,290
What does that mean?

32
00:01:43,290 --> 00:01:46,380
The ramen there, forgettable.

33
00:01:46,380 --> 00:01:46,900
Don't want to eat it.

34
00:01:47,920 --> 00:01:49,030
But I'm not going to restaurant for

35
00:01:49,030 --> 00:01:51,920
ramen, I don't care about
this sentence in the review.

36
00:01:51,920 --> 00:01:55,630
On the other hand, the same restaurant
might say, the sushi there was delicious,

37
00:01:55,630 --> 00:01:57,740
it was the best sushi in Seattle.

38
00:01:57,740 --> 00:01:59,070
Now that says a lot about the sushi,

39
00:01:59,070 --> 00:02:04,110
which is the thing I care about the most
with respect to that restaurant I choose.

40
00:02:04,110 --> 00:02:06,640
So every review has different aspects and
different sentences.

41
00:02:06,640 --> 00:02:11,000
I would like capture a sentiment of each
sentence so I can understand if it's good

42
00:02:11,000 --> 00:02:13,190
with respect to sushi which is
what I care the most about.

43
00:02:14,190 --> 00:02:16,830
So every sentence,
I'm going to feed it to a classifier.

44
00:02:16,830 --> 00:02:21,440
It's going to say, is this a positive
sentiment or is this a negative sentiment?

45
00:02:21,440 --> 00:02:22,561
And that's our task.

46
00:02:22,561 --> 00:02:26,939
[MUSIC]