Altavista - Intro to Computer Science Udacity English Japanese

Altavista - Intro to Computer ScienceYouTube↗ IA↗(long shot)
Education 3.9K views 6 likes 6 dislikes 23 Feb 2015

Udacity 651K subscribers United States YouTube↗

Click to Play Video

So the final change we're going to make, is thinking about our

私たちが行おうとする最後の変更はランダムサーファモデルについて考えることです

random surfer model. If this was our model, if a page has

もしこれが私たちのモデルだとするとどこからもリンクが張られていないページは

no incoming links, well, it's popularity ranking would be 0. We

人気ランキングがゼロになるでしょう

don't want to have that be the case. If we think of

そのようなケースは持ちたくありません

our random web surfer, well if she randomly starts at a

ランダムWebサーファについて考える時サーファがランダムに開始ページを決めるならば

page, she could randomly start at a page that has no links.

リンクのないページからランダムに開始することがあるでしょう

If we made all the page scores 0, for pages with no

リンクのないページの人気度をすべてゼロにすると

links, well, then it would be very hard for a new page

新しいページを開始するのが非常に難しくなります

to get started. So, we don't want those to be 0,

よってその人気度をゼロにしたくないので

so we're going to have some random probability that you reach a page even when it has no links. So that means we're going to

リンクがないページにもたどり着くランダムな可能性を持つようにします

add something to this sum so it's not zero when there are

つまりこの合計に何かを加えて

no incoming links. And the other thing we're going to do

受けるリンクがなくても人気度がゼロにならないようにします他に私たちが行おうとしているのは

is we're going to scale this summation a little bit. And we're

この合計を定めることです

going to scale that with what is called the damping function. And

ここでは減衰関数と呼ばれるもので定めます

that just means, if we think of our random web surfer again,

その意味を説明しましょうランダムWebサーファについて再び考えると

even if she's following a path. That does have more

ランダムWebサーファの経路に別のリンクがある場合でも

links. At some point, she might decide to get tired and

ある時点でサーファは疲れてしまい

give up and start again with a new random page. So,

リンクをたどるのをやめて新しいランダムページから始めるかもしれません

the dampening function determines how frequently, we think our random web

そこで減衰関数が決定するのは

surfer will pick a random link, versus starting over again

ランダムWebサーファがどのくらいの頻度で新しいランダムページから始めるよりもランダムリンクを選ぶかということです

on a new random page. So that's call the dampening value

それが減衰値と呼ばれるものです

and we are just going to use the dampening value to scale

この減衰値を定めたいと思います

this number. So, we'll call that d, that's the dampening constant.

それをdとしましょう　減衰定数です

A typical value for that is something less than one,

減衰定数の典型的な値は1未満であり

and a good value is something like 0.8. So now we're

よい値は0．8くらいです

going to change our rank function to take into account that. We're

そのことを考慮に入れるために rank関数を変更しましょう

going to have some initial value and if we want to keep the

何らかの初期値を持たせます

values in a reasonable range, instead of starting from one,

その値を合理的な幅に収めたいならば 1から開始するのではなく

we're going to divide the values by a total number of pages.

その値をページの合計数で割ります

And the reason for that is just to keep the ranks in some meaningful range. This will keep the ranks so the total of the ranks when we start the sum of the

ランクを意味のある幅の中に収めておくためですこうすればランク合計を始めた時のランク合計値が1に保たれます

ranks is 1, that gives them a little more meaning then if the sum of the ranks was the number of pages. So

この方がランクの合計がページ数になるよりも意味のあるものとなります

n is going to be the number of pages, the total number

Nはページ数つまり私たちのコーパスにおける合計数です

in our corpus and d is the damping constant, and so

dは減衰定数です

now we're going to change our initial values to divide them by n.

ここで初期値をNで割って値を変更します

This means, the sum of the paid ranks at the beginning will

これは当初のページランクの合計が 1になるという意味です

be 1, and, we're going to change the way we compute the rank

そしてランクを計算する方法を変えましょう

by adding to it, a value that correspo... That gives us the sense of starting over. So, the value that we're

ここにやり直すという意味になる値を加えます

going to use for this is 1 minus d, divided by n

その値としては（1－d）／Nを使います

and, that gives the notion that. That times when we don't decide to do a new page. So, 0.8 of the time. We decide to follow a new page. So we're going to

そして新しいページを開始しない場合すなわちリンクをたどる確率dは 0．8と定めましょう

multiply this by D. The times when we don't follow a new page, that's one minus D. We're going to start over, and

これにdを乗じますリンクをたどらない時すなわち1－dの場合はやり直しとなりこの値をNで割ります

we're going to divide that by N, because N is the number of pages that the probability that this is the page that would be picked when we start over. So now we've

なぜならページ数のNで割るとやり直した場合に選ばれる可能性を表せるからです

got our recursive definition of page rank. We start by initializing the rank n times at 0 to one minus, divided by

これでページランクの再帰定義ができましたまずrankの時間ステップがゼロの時の初期値として 1／Nを設定します

N, and then for as many times up, as we want,

そして任意の時間ステップについて

we're going to keep improving our results. By using this formula. For

この公式を使って結果を改善し続けます

each link that links to us, we're going to take it's popularity

こちらにリンクを張っている各ページについて前回のステップでのそれぞれのランクを使います

on the last step, divide that by the number of out links so it's dividing it by the number of other pages it

その際にはそのページから出ている総リンク数でランクの値を割ります

links to. Multiply that by our damping constant and that gives

それに私たちの減衰定数を乗じると

the probability that this page was selected by the random web surfer,

このページがランダムWebサーファに選ばれるという可能性になります

and then we're going to add the term that takes into

そこに追加する項では

account the random web surfer might have started over from scratch and

ランダムWebサーファが初めからやり直して

picked a new random page. So, before we try to change our Web search engine to actually program this, it's time for a

新たなランダムページを選ぶ可能性を考慮しますこのプログラムを実際に使うため私たちのWeb検索エンジンを変えたいと思いますが

little quiz. And the quiz is, What is AltaVista? And if you don't know this,

その前に小テストを行います AltaVistaとは何でしょうか？

feel free to use Google, or DuckDuckGo, to find out. These are the choices so,

知らなければGoogleや DuckDuckGoで調べていただいて構いません選択肢はこちらです

it's the view from the Udacity headquarters

パロアルトにあるUDACITY本部からの眺め

in Palo Alto. It's the most popular web search engine in 1998. It is Spanish for, You're Terminated Baby and, it's a small town

1998年に最も普及していた検索エンジンスペイン語で“もう終わりだ！”という意味バージニア州の小さな町

in Virginia. And feel free to Google the answer if you don't know.

分からなければGoogleで検索してみてください

Similar videos:

Same Value - Intro to Computer Science
Udacity
21.5K views 6 likes 0 dislikes 23 Feb 2015100% match

Testing2 - Intro to Computer Science
Udacity
19.4K views 2 likes 0 dislikes 23 Feb 2015100% match

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

Find Element - Intro to Computer Science
Udacity
22.1K views 19 likes 0 dislikes 29 May 2012100% match

Test for Valid Inputs - Intro to Computer Science
Udacity
22.1K views 1 likes 5 dislikes 23 Feb 2015100% match

First Step - Intro to Computer Science
Udacity
29K views 5 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

Introduction
Udacity
87.8K views 315 likes 4 dislikes 25 Apr 2014100% match

Strings Solution - Intro to Computer Science
Udacity
19.4K views 22 likes 6 dislikes 02 Mar 2012100% match

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

Abbaize - Intro to Computer Science
Udacity
30.6K views 32 likes 0 dislikes 02 Mar 2012100% match

Varying Variables Quiz 2 Solution - Intro to Computer Science
Udacity
19K views 5 likes 0 dislikes 23 Feb 2015100% 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

Page Rank - Intro to Computer Science
Udacity
11.6K views 71 likes 1 dislikes 23 Feb 2015100% match

Speed of Light - Intro to Computer Science
Udacity
67.4K views 17 likes 0 dislikes 23 Feb 2015100% match

More Friends - Intro to Computer Science
Udacity
28.1K views 31 likes 0 dislikes 02 Mar 2012100% match

Strings - Intro to Computer Science
Udacity
27.1K views 22 likes 0 dislikes 02 Mar 2012100% match

Define daysBetweenDates - Intro to Computer Science
Udacity
23.7K views 5 likes 0 dislikes 23 Feb 2015100% match

How Computers Store Data - Intro to Computer Science
Udacity
40.1K views 158 likes 55 dislikes 29 May 2012100% match

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

Varying Variables 1
Udacity
35.1K views 4 likes 0 dislikes 23 Feb 2015100% match

Sum List - Intro to Computer Science
Udacity
20.5K views 15 likes 1 dislikes 29 May 2012100% match

Biggest - Intro to Computer Science
Udacity
30.2K views 28 likes 0 dislikes 02 Mar 201278% match

Testing - Intro to Computer Science
Udacity
31.4K views 6 likes 0 dislikes 23 Feb 201578% match

Median - Intro to Computer Science
Udacity
30K views 21 likes 0 dislikes 23 Feb 201578% match

Square - Intro to Computer Science
Udacity
32.5K views 32 likes 0 dislikes 02 Mar 201278% match

Udacify - Intro to Computer Science
Udacity
26.4K views 17 likes 0 dislikes 23 Feb 201578% match

Hello!!! - Intro to Computer Science
Udacity
51K views 7 likes 0 dislikes 23 Feb 201560% match

Settings Privacy About Contact Donate