Page Rank - Intro to Computer Science Udacity English Chinese (China)

Page Rank - Intro to Computer ScienceYouTube↗ IA↗(long shot)
Education 11.6K views 71 likes 1 dislikes 23 Feb 2015

Udacity 651K subscribers United States YouTube↗

Click to Play Video

So, that idea for how to rank webpages is

对网页进行排名与

the same idea as how we measure popularity of people.

衡量人们的受欢迎程度相似

But instead of thinking about friendships as the way

我们要衡量的是网页上的链接

to measure popularity, what we're measuring is links on the

而不是好友关系

web. When one page has a link to another

当某个网页有一个指向其他网页的链接时

page, well that indicates it's more likely that this other

则表明这个其他网页很有可能比较热门

page is popular, just like when someone is a friend,

就像某个人是好友一样

it indicates that the person they are friends with is

表明某人的好友更有可能比较受欢迎

more likely to be popular. So, the goal in ranking web

网页排名的目标是

pages is to get a measure of how popular are pages,

根据链接到相关网页的其他网页

based on the other pages that link to it but we

衡量该网页的热门程度

have the same issue with popularity then not all links are

但是我们遇到了和受欢迎程度一样的问题

the same. That a link from a page that's really important

并非所有的链接都一样某个非常重要的网页发出的链接

counts for a lot more than a link from a page

比不重要的网页发出的链接占比要大很多

that is not important. So, if the New York Times has

假设《纽约时报》上有个链接指向你的网页

a link to your page, that counts for a lot more than

同时假设你妈妈设置了一个网页

if your mom sets up a web page and puts a

并在上面放置了指向你的网页的链接

link to your page in it. Unless your mom is Lady

那么纽约时报的链接肯定更具分量

Gaga, in which case her link probably counts for more. So,

除非你妈妈是 Lady Gaga 这种情况下她给出的链接更具分量

another way of thinking about this is what we're trying to model

另一种思考方法是我们要模拟

is a random web surfer. So, our random web surfer has

一个随机网络服务器

some set of pages that they know about. And those pages

该服务器具有一些已知的网页

have links to other pages. Some pages might have a lot

这些网页具有指向其他网页的链接

of links. Some pages might just have one link. Some pages might

某些网页可能具有很多链接某些可能只有一个链接

have no links. So, one way to think about this

某些可能没有链接

is that we're trying to model a random web surfer. The

一种思考方式是尝试模拟一个随机网络服务器

web surfer starts knowing about some pages. And she picks

该服务器开始了解一些网页

one page at random, let's say she picks this page. And

它随机选择一个网页假设选择了这个网页

then, when she's on that page, she picks a random

然后转到该网页上随机选择一个链接

link and follows that links. Whoops, this was a bad starting

点击该链接糟糕这个起始网页不太好

page. It actually has no out links. So, then, she picks

它实际上没有向外的链接

another random page. Let's say she picks this one. She follows

然后选择另一个随机网页假设选择了这个

the link from that page, and now she got to the

点击该链接

page with no links again. Let's say she picked a better starting

再次转到没有链接的网页假设选择了更好的起始点

point. Let's say she randomly picked this one. Now she's got

假设随机选择了这个网页

two links to follow. She randomly picks one of those. She follows

现在可以点击两个链接随机选择一个

it. She gets to a new page. She randomly picks a

转到新的网页

link from that page, in this case, it only had one, and

随机选择该网页上的链接这种情况下只有一个链接

in this case, it seems we have a bit of a

在这种情况下我们似乎遇到了问题

problem, because all of the starting pages eventually lead into this one,

因为所有的起始网页最终都转到这个网页

which has no outgoing links. So, we'll think about how to

它没有向外的链接

solve that problem later, but we can think of what our

稍后我们将思考如何解决该问题

random web surfer is doing, is picking random pages, following links,

现在我们可以认为随机网络服务器所做的是选择随机网页

and what we want to measure is the popularity of a page.

点击链接我们想要衡量某个网页的受欢迎程度

And that's the probability that she reaches that given page, starting

即从这些随机网页抵达给定网页的概率

from these random pages. So, if you did this over and

如果不断执行这一流程

over again, and you counted the number of times you reached

算出抵达每个网页的次数

each page, that would give you a measure of that page's popularity.

就可以得出该网页的受欢迎程度

So, this is very similar to the popularity function. We're

和 popularity 函数很相似

going to define a function that we'll call the rank of

我们将定义一个网页排名函数

the page. And, like the way we defined the popularity

和定义 popularity 函数一样

function, it's going to have two inputs. It's going to have a

有两个输入

time stamp, and it's going to have the page which we'll

一个是时间步长一个是网页用 URL 表示

use a URL for. And the output of rank will

rank() 的输出是一个数字

be some number. Except for we'll define for time step

但是我们将定义时间步长 0

zero. This is our base case, and we're going to find all

这是我们的基础条件

the ranks have value one. We'll actually change

所有排名值为 1

that shortly, but we'll start out think of it,

稍后做更改我们先假设所有排名

all the ranks having value one, like we

did with the popularity scores. And then we'll define

就像受欢迎程度得分一样

the value of the rank at time step t for any given URL. Just like we defined

然后将定义任何给定 URL 在时间步长 t 时的排名值

the popularity score. It's going to be the sum of

就像定义受欢迎程度得分一样

all the pages that are friends with this page,

等于与该网页是好友的网页排名之和

and what it means for our webpage to be friends

网页与其他网页是好友是指

with another page is that it has a link to it.

具有指向该网页的链接

So, this is going to be for all the pages that

也就是存在的链接到该 URL 的所有网页

exist that have some link to that URL, or its friends.

即它的好友

And so, we're going to go through each of those pages.

我们可以查看每个网页

We'll call them inlinks instead of friends. We're going to go through

称之为 inlinks 而不是 friends

those and we're going to sum up the ranks that we got

我们将查看这些网页并算出时间步长 t-1 时

for those pages at time t minus one. So, this is

这些网页的排名之和

our first model of popularity on, of webpages. This is exactly

这就是网页的第一个受欢迎程度模型

the same as the model we had of popularity for people.

这和人的受欢迎程度模型完全一样

It's not going to work that well yet and one of the

但是效果不太好

reasons it's not going to work that well is some pages

原因之一是某些网页

might have lots of links. And if a page has lots

可能有很多的链接

of links, the value of each one of its links should

如果某个网页有很多的链接则它的链接的每个值

be diminished. It shouldn't have the same value as the page

都应该减小不应该和只有一个链接

that only has one link that links to this URL. Maybe that

指向该 URL 的网页值相同

should be the same case for a friend's popularity, if

或许这和好友的受欢迎程度差不多

someone has lots of friends, each friend is less valuable.

如果某人有很多好友则每个好友的重要性都更低

Whereas someone who only has a small number of friends

但是如果某人只有一小部分好友

has lots of time for each friend. So this is the

则有很多时间陪每位好友

way we're going to model web popularity. We don't want

我们就这样表示网页受欢迎程度

to just give the same score to each link. We're

我们不希望给每个链接设定相同的得分

going to change this by dividing by the number of outlinks.

为此我们将除以向外链接的数量

So, if a page has many outgoing links, the value to

如果某个网页有很多向外链接

the pages that it links to is less for each page. So,

则它链接到的每个网页的值会降低

a page that's just a long list of lots of links won't

如果某个网页是一个长长的链接列表

have that much influence on the rankings. If a page only has

则对排名没太多影响

a few outgoing links, well then, they are worth more. So, what we are

如果某个网页只有少数几个向外链接那么重要性更高

going to do, is divide this by the number of out links from

我们要做的是将这个除以 p 的向外链接数量

p. So, each of the p pages, right? These will be the

即每个 p 网页对吧？

values of p, as they go through the inlinks of URL. We

这些将是经历 URL 的内部链接时 p 的值

are going to sum up the rank that we got on the previous time

我们将对上一时间步长的排名求和

step and divide that by the number of out links.

并除以向外链接的数量

Similar videos:

Best Strategy - Intro to Computer Science
Udacity
20K views 2 likes 3 dislikes 23 Feb 2015100% match

What Are the Outputs
Udacity
27.5K views 5 likes 0 dislikes 23 Feb 2015100% match

First Programming Quiz - Intro to Computer Science
Udacity
97.3K views 56 likes 3 dislikes 23 Feb 2015100% match

Valid Strings Solution - Intro to Computer Science
Udacity
14.8K views 7 likes 0 dislikes 23 Feb 2015100% match

Define nextDay - Intro to Computer Science
Udacity
21.7K views 2 likes 0 dislikes 23 Feb 2015100% match

Finding with Numbers Quiz - Intro to Computer Science
Udacity
32.5K views 4 likes 0 dislikes 23 Feb 2015100% match

巴克斯构造
 Udacity
55.9K views 356 likes 12 dislikes 19 Feb 2012100% match

Valid Strings - Intro to Computer Science
Udacity
35.1K views 3 likes 0 dislikes 23 Feb 2015100% match

Finishing the Web Crawler - Intro to Computer Science
Udacity
13.4K views 22 likes 0 dislikes 29 May 2012100% match

Capital Udacity - Intro to Computer Science
Udacity
54.1K views 8 likes 1 dislikes 23 Feb 2015100% match

Is Friend - Intro to Computer Science
Udacity
32.1K views 26 likes 0 dislikes 02 Mar 2012100% match

Countries - Intro to Computer Science
Udacity
20K views 23 likes 0 dislikes 29 May 2012100% match

Index - Intro to Computer Science
Udacity
26.9K views 37 likes 1 dislikes 29 May 2012100% match

Python Expressions - Intro to Computer Science
Udacity
64K views 39 likes 4 dislikes 23 Feb 2015100% match

Procedure Code - Intro to Computer Science
Udacity
23.4K views 34 likes 4 dislikes 02 Mar 2012100% match

What Should We Write First - Intro to Computer Science
Udacity
12.2K views 1 likes 3 dislikes 23 Feb 2015100% match

Biweekly Solution - Intro to Computer Science
Udacity
27.5K views 40 likes 0 dislikes 23 Feb 2015100% match

Understanding Selection Solution - Intro to Computer Science
Udacity
19.2K views 19 likes 1 dislikes 23 Feb 2015100% match

Should We Implement It - Intro to Computer Science
Udacity
12.1K views 1 likes 1 dislikes 23 Feb 2015100% match

First Step - Intro to Computer Science
Udacity
11.4K views 4 likes 0 dislikes 23 Feb 2015100% match

Bodacious Udacity - Intro to Computer Science
Udacity
41.1K views 20 likes 0 dislikes 02 Mar 2012100% match

Testing2 Solution - Intro to Computer Science
Udacity
18.9K views 17 likes 3 dislikes 23 Feb 2015100% match

Eat Quiz Solution - Intro to Computer Science
Udacity
21K views 16 likes 0 dislikes 23 Feb 2015100% match

Same Value Solution - Intro to Computer Science
Udacity
22.3K views 18 likes 6 dislikes 23 Feb 2015100% match

What Is a Program?
Udacity
100.9K views 32 likes 5 dislikes 23 Feb 2015100% match

Return Statement - Intro to Computer Science
Udacity
18.3K views 11 likes 4 dislikes 23 Feb 2015100% match

Median - Intro to Computer Science
Udacity
30K views 21 likes 0 dislikes 23 Feb 2015100% match

Find Second - Intro to Computer Science
Udacity
20.2K views 33 likes 0 dislikes 02 Mar 2012100% match

First Quiz - Intro to Computer Science
Udacity
48.4K views 28 likes 5 dislikes 23 Feb 2015100% match

Days in a Month - Intro to Computer Science
Udacity
20.7K views 14 likes 1 dislikes 29 May 2012100% match

Settings Privacy About Contact Donate