SEO 通过 的 Sea ⚓

Learn SEO Directly from 的 Search Engines

Is This Really 的 Panda Patent?

分享是关怀!

谷歌是否’Navneet Panda共同发明的一项新专利描述了Google’s Panda Update?

搜索质量与网络垃圾邮件

我的许多专利申请’ve written about from 谷歌 address webspam. They also look at 怎么样 search engines may follow approaches to keep search results from manipulation. An early example 是 的 patent from 2003 titled 识别被操纵物品的方法和系统.

Is this 谷歌's Panda?

But many of 的 patents I’ve written about involve 谷歌 trying to improve 的 质量 of search results for searchers.

谷歌之一’的第一项专利(记住PageRank是斯坦福大学’而不是Google的专利’s)涉及查看响应于查询的最高搜索结果,并提高该查询的某些搜索排名(如果它们与同一查询的其他排名靠前的链接)。该专利, 排名搜寻结果 通过 reranking 的 results based on local inter-connectivity, 原为 aimed at improving 的 质量 of 的 top-ranking results.

谷歌’s 基于短语的索引 patents involve meaningful words and phrases that tend to co-occur in 的 search results for a specific query. 的rankings of those Pages could be boosted when those phrases do appear 要么 boosted 通过 怎么样 much weight 是 passed along through anchor text using one of those related co-occurring terms. These are search 质量 patents.

有几种基于短语的索引专利,并且’通过检查页面上的结果是否存在统计上异常数量的同时出现的单词,这也是解决Web垃圾邮件的至少之一。因此,基于短语的索引方法还包括一种检测Web垃圾邮件的方法。

注重质量

授予Navneet Panda和Vladimir Ofitserov的专利, 排名搜寻结果旨在改善搜索结果,而不是惩罚网站或确定操纵搜索结果的尝试。

的patent does list 只要 one “advantage” to following 的 process it describes:

Search results identifying low-quality 资源s can be demoted in a presentation 要么der of search results returned in response 到 用户’的查询。因此,由于呈现顺序较高的搜索结果将更好地匹配用户,因此可以改善用户体验。’的信息需求。

Before 的 Panda update, 的re had been many public criticisms of 的 质量 of search results showing up in searches at 谷歌.

这里 are a few examples:

2009年12月13日–洗碗机以及Google如何吃自己的尾巴– Paul Kedrosky

谷歌 has become a 太容易消耗自己的关键字tail的蛇。识别一些在获利搜索中显示的单词“从设备到间皮瘤套装,再到皮划艇课程,廉价而定期地制作出内容,您’re done. On 的 web, no-one knows you’是一个内容研磨机。

2009年12月13日– 内容农场:为什么选择媒体,博客&谷歌应该担心 – Richard MacManus

根据我对Demand Media和类似网站的分析,此类内容非常通用且缺乏深度。虽然我不会’一直到WikiHow创始人Jack Herrick说“lacks soul,”当然,它缺乏热情,经常也缺乏对当前主题的了解。阿灵顿’s 与快餐类比是恰当的–它是快速生产并按订单生产的内容。

2011年1月2日– On 的 increasing uselessness of 谷歌….. – Alan Patrick

但是今年它受到了谷歌的打击’的系统已被发送垃圾邮件,因为 typically anything on Page 1 of 的 search results 原为 some form of SEO spam –最典型的是’不会卖给您任何东西,只需指向其他网站(通常做同样的事情),同时向您滑动一些广告(毫无疑问是“relevant”). 的other main scam 现场 type 是 one that copies part of 的 相关的 Wikipedia entry and throws lots of Ads at you

2011年1月3日– Trouble In 的 House of 谷歌 – Jeff Atwood

像任何理智的人一样,我’我在这场战斗中扎根Google,而我’d love 不hing more than for 谷歌 to tweak a few algorithmic knobs and make this entire blog entry moot. Still, this 是 的 first time since 2000 that I can recall 谷歌 search 质量 ever 下降并激发了我一些异端思想— are we seeing 的 first signs that algorithmic search has failed as a strategy? Is 的 next generation of search destined to be less algorithmic and more social?

It’甚至是一件令人恐惧的事情, 但是也许引力真的被打破了。

2011年1月27日– 谷歌 Search Quality Decline 要么 Elitism? – AJ科恩

谷歌当然可以做到这一点。他们可以站起来说,Demand Media提供的快餐食品不会’获得优质的SERP房地产。 谷歌可能会进行优化,而不是优化。 他们可以选择快餐以外的美食。

But 是 that what 的 ‘user’ wants?

提高品质

As you see from those quotes, 的re 原为 a sense of 谷歌 results being broken, and showing results more focused upon matching queries than returning 质量 results.

These criticisms were heard, even at 的 谷歌plex. In February of 2011, 的 Official 谷歌 博客 told us of an update in 在搜索中找到更多高质量的网站. This change covered a fair number of searches, and 原为 aimed at surfacing higher 质量 content:

但是在最后一天左右,我们对排名进行了很大的算法改进,“这一变化明显影响了我们查询的11.8%”我们想让人们知道’继续。此更新旨在降低对用户的低附加值的“低质量网站”的排名,从其他网站或不太有用的网站复制内容。同时,它将为具有原始内容和信息(例如研究,深入报告,周到的分析等)的高质量网站提供更好的排名。

My question, after watching 的 Panda update and reading a lot of threads in forums and other places about 现场s impacted 通过 Panda, and working on 现场s that were, 是 whether 的 patent from Navneet Panda describes 的 update, and attempts to improve 的 质量 of search results.

这里’s a quick summary from 的 patent of what happens in 的 process it describes:

  • Determining, for each of a plurality of 团体 of 资源s, a respective count of 独立 incoming links to 资源s in 的 组;
  • Determining, for each of 的 plurality of 团体 of 资源s, a respective count of 参考查询;
  • Determining, for each of 的 plurality of 团体 of 资源s, a respective 组-specific modification factor, wherein 的 组-specific modification factor for each 组 是 based on 的 count of 独立 links and 的 count of 参考查询 for 的 组; and
  • Associating, with each of 的 plurality of 团体 of 资源s, 的 respective 组-specific modification factor for 的 组, wherein 的 respective 组-specific modification for 的 组 modifies initial scores generated for 资源s in 的 组 in response to received search queries.

So 的 patent has multiple 部分, which work together.

的first involves looking at 的 links pointing to 的 页s of a 现场, and removing all of 的 backlinks that look like 的y might be affiliated (under co-ownership 要么 control) with 的 现场, 要么 reducing 的 number of 独立 links 到ccount for things like 现场-wide links. This may be done to get a sense of 怎么样 many different unrelated 页s and 现场s are linking to 的 页s of this 现场, 的 more 独立 links from more sources could be seen as a sign of 质量.

第二个是分析页面是否似乎针对特定的引荐查询。虽然它’对于对网站进行SEO尝试使网站上的每个页面都成为潜在目标页面的人来说,这并不罕见,我们称为内容农场网站的许多网站经常使用每个页面来定位高度商业化的查询,以及针对这些查询的多种变体查询。因此,内容服务器场类型的网站可能包含许多试图引用很多查询的页面。

的独立 links count and 的 参考ring queries count for 的 different 团体 that a 现场 might be broken down into, are looked at as a ratio, with 独立 link count over 参考ring query count. If 的re are a lot of 独立 links and few 参考ring queries, this number could be over one. If 的re are few 独立 links and lots of 参考ring queries, 的 number could be a fraction of one.

This number based upon 的 links and 的 queries 将 的n be multiplied 通过 a score that has been modified 通过 whether 要么 不 each 页 是 seen as a 航海的 type result for a query term 要么 phrase. 的more it 是 like a 航海的 term 要么 phrase, 的 higher this part of 的 score. 的final score could boost ranking scores for some results and diminish scores for other results.

组s Rather Than Pages

该专利没有像许多排名算法那样针对特定页面或网站,而是告诉我们它关注的是“groups” of 资源s. A 组 might be defined in several different ways. 资源资源 within 团体 can 只要 be included in one 组.

A 组 might be address based, so that all of 的 资源s within 的 组 are all in 的 same domain name, such as http://www.example.com. Or all in 的 same hostname on a domain, such as http://host1.example.com 要么 http://host2.example.com.

组s of 资源s might be partitioned 通过 a count of 参考查询 for each of 的 团体, “so that each partition includes 团体 of 资源s whose counts of 参考查询 are within a respective range of counts of 参考查询.”

Under this approach, one 网站 might be broken into more than one 组 要么 might be part of a 组 that contains more than one web 现场. To rank 的 页s within 的se 团体, 的 ratio of 独立 links to 参考查询 might be multiplied 通过 a score involving 航海的 signals to determine a final rank.

独立链接

If 的 purpose behind this patent 是 to rank 页s higher that are higher 质量, one way to do that 是 to look at 的 number of 独立 links to those 页s, 要么 团体 of 页s.

For each of 的se 团体 of 资源s, 的 patent tells us that it might count 的 number of links to those 团体 –但并非所有链接。并不一定表示链接– links that you can click upon to get 到nother 页. It might also count 默示 links, which sound more like what we often tend to 参考 到s citations. An express link can be used to navigate 到 place, where an 默示 link can’点击以将一个人带到该链接的目标。

为什么不’t 的 patent mention PageRank? Both this metric and PageRank are supposed to be signals of 质量, but 不 every signal from 谷歌 has to include PageRank. This reliance on 独立 links eliminates 的 benefit of having a 现场 with lots and lots of 页s to be linked to from 的 same 现场 要么 现场s that are under 的 same ownership 要么 control. Or linked to 现场wide from other 现场s.

An 独立 link 是 where 的 source of 的 link and 的 target 是 determined to be 独立 of each other. 的source 组 that a link 是 in and 的 target 组 can be checked to see if 的y are 独立 of each other as well.

Determining that links from one 组 到nother are 不 独立 links can also involve determining that those 团体 of 资源s are likely to be related, such as owned, hosted, 要么 created 通过 的 same entity.

If 的 资源s have similar 要么 identical content, 要么 images, 要么 formatting, 要么 CSS 要么 so on, 的ir similarity 是 another sign that 的 资源s are 不 独立.

Where 的re may be multiple links from one 资源 到 targeted 组, 只要 one of 的 links might be counted as an 独立 link. Though it’s 不 said in 的 patent, 这样可以避免站点范围内的链接被多次计数。

参考查询

In addition 到n analysis of 的 links pointing to 的 different 参考ence 团体, this process also looks at 的 页s of 的 现场, and 的 queries that each might be targeting. How well do those 页s satisfy those queries?

如果页面包含该词“example.com”, it might be said to 参考 to 的 home 页 of a 现场. If it includes terms that are commonly used 通过 searchers to 参考 to 的 页s of a 现场, it might be said to include 参考ring queries that 参考 to those 页s. 的patent provides an example of others, 通过 telling us that:

…if 的 terms “example sf” and “esf” are often used 通过 searchers to 参考 to 的 资源 whose URL 是 “http://www.sf.example.com,” queries that contain 的 terms “example sf” 要么 “esf”, e.g., 的 queries “example sf news” and “esf餐厅点评,” can be counted as 参考查询 for 的 组 that includes 的 资源 whose URL 是 “http://www.sf.example.com.”

导航查询

In 的 post, 谷歌如何识别导航查询和资源,我写了一篇文章,介绍了Google如何使用文档分类方法来确定搜索者是否输入了要查询的页面,并希望找到特定页面,例如查询中包含的产品或服务的官方主页。

在某种程度上,这种询问是’t too different than 的 set of questions that are raised in Amit Singhal’在Google博客的官方帖子中, 有关建设高质量网站的更多指导. Such questions were worked into an analysis of a 现场 at a stage like at this point, though 的 patent doesn’t 参考 to 的m specifically.

的patent 是:

排名搜寻结果
由Navneet Panda和Vladimir Ofitserov发明
Assigned to 谷歌
美国专利8,682,892
授予:2014年3月25日
提交日期:2012年9月28日

抽象

用于对搜索结果进行排名的方法,系统和装置,包括在计算机存储介质上编码的计算机程序。

One of 的 methods includes:

  • Determining, for each of a plurality of 团体 of 资源s, a respective count of 独立 incoming links to 资源s in 的 组;
  • Determining, for each of 的 plurality of 团体 of 资源s, a respective count of 参考查询;
  • Determining, for each of 的 plurality of 团体 of 资源s, a respective 组-specific modification factor, wherein 的 组-specific modification factor for each 组 是 based on 的 count of 独立 links and 的 count of 参考查询 for 的 组; and
  • Associating, with each of 的 plurality of 团体 of 资源s, 的 respective 组-specific modification factor for 的 组, wherein 的 respective 组-specific modification for 的 组 modifies initial scores generated for 资源s in 的 组 in response to received search queries.

观察结果

Chances are that 谷歌 tweaked and changed 的 Panda Algorithm in 的 weeks and months after it 原为 first applied, and may have made many changes to it as it went through an initial beta period.

I’ve seen several denials from people about this particular patent describing 的 Panda update since I wrote about finding it last week in 谷歌的Panda授予了对搜索结果排名的专利。这些拒绝是基于专利中描述的链接分析的存在,而没有更仔细地研究所涉及的实际过程,并声称该专利比熊猫方法更可能详述Penquin方法。

But 的 link analysis here involving 独立 links, and 参考ring queries are more of an attempt to gauge 的 质量 of a 现场 than 的 backlink profile of that 现场. 的“navigational” query analysis that could involve 是sues such as 的 example 23 questions that Amit Singhal provided us with also attempts to understand 的 质量 of 页s.

I ask, with 的 title of this post, if this patent 是 “really 的 Panda Patent”, but I do think it 是. But I am open to 的 possibility that 的 Panda Updates followed a somewhat different course as 的y were implemented and tested.

分享是关怀!

关于61条想法“Is This Really 的 Panda Patent?”

  1. Thanks for sharing for such a detail 文章 Panda patent, Bill. 您 have input such great analysis into this blog post. Thanks again for 的 great share of your thoughts!

  2. 非常有趣的比尔。

    的first few 部分 of 的 patent, as you describe 的m, seemed aimed at Demand Media. Removing ‘controlled’ links (from other properties under corporate ownership) and 的n 的 links to 参考ring queries ratio, which 将 target those 现场s that were ‘sharding’ keywords (i.e. –如何烧开水,如何烧开热水,如何烧开冷水)。

    And 的 团体 of 资源s certainly dovetails with Panda too, since we all saw that it 原为 generally a 现场 based demotion and 不 done on 的 document level.

    I’m guessing this 原为 a material part of 的 Panda update. And it 将 make sense since it 原为 undertaken because of 的 chatter you 参考ence. Hence 的y started with a target 组 –可以这么说学习集。

    Not 只要 that, but it underscores 的 fact that links were (and still are) an important way to determine 质量. That doesn’t表示PageRank,但链接图分析简单。

  3. 嗨,比尔,

    很棒的分析!

    一个小细节引起了我的注意,所以我’您已经在您的官方文档中对其进行了检查’ve mentioned – in regard to 参考查询 的 patent quotes: “For example, a term that 参考s 到 资源 may be all of 要么 a portion of a 资源 identifier, e.g., 的 URL, for 的 资源.” But 是n’与Google所说的关键字填充域问题相矛盾,该声明降低了谷值作为排名因素。

    For instance if you are reaching 到 given 资源 thanks 到 keyword query and it matches 的 domain of a given 相关的 网站 (with keyword stuffed url) 的n it should 不 be taken so much into account as to boost 的 given 网站 in 的 SERPs …或我缺少一些东西。无论如何–完美的解释,做得很好:)

  4. Stellar stuff as 总是 Bill.

    Like AJ I find 的 “群组而不是页面”部分可能会引人注目的熊猫爪印。不仅是因为网站受到了打击,还因为域名受到了严重打击 部分 一个网站遭受的打击比其他网站严重。

    在定位的网站内容中,这很有意义“高度商业化的查询以及这些查询的多种变体。”人们通常会在内容农场中找到的东西可能被描述为“目标查询页面集群” which, ironically, exposed 的mselves for classification as low 质量 通过 virtue of 的 nature of 的ir targeting. Pretty clever of 谷歌 to leverage this particular signal.

  5. 嗨,比尔,

    Illuminating post. Could you clarify 的 meaning of ‘reference queries’ for me in 的 context of “Determining, for each of 的 plurality of 团体 of 资源s, a respective count of 参考查询”.

    I’m surprised to see you 参考 to PageRank as a 质量 signal, instead of a popularity signal. I’m guessing you mean that 质量 是 inferred 通过 popularity.

    I agree with Michael M that 的re must be many other patents that cumulatively protect Panda.

    BTW, I like 的 现场’新的标题图片和设计。

  6. Hi 戴夫 ,

    鉴于最近几年美国专利法的变化以及对现有技术的变化,我不’不能肯定。熊猫更新的使用是否早于本专利的提交及其使用方式是否已使其算作现有技术?我不是专利律师,我可以’给您一个清晰的答案。

    的claims within 的 patent, and 的 description do seem like 的y 将 fit 的 Panda algorithm, though 的 algorithm 是 one that appears to have been evolving over time as well, and this patent might cover a process that 原为 adopted along 的 path of its development 而不是at 的 start of it.

    而且,正如迈克尔·马丁内斯(Michael Martinez)先前指出的,’t know if 的re might 要么 might 不 be other patents that are possibly related that have been filed as well. If so, 的y could cover 部分 of 的 Panda algorithm as well.

  7. 嗨,AJ,

    该专利的确似乎是用Google的某些批评中所讨论的各种网站编写的’的搜索结果,并且所描述的方法似乎确实解决了我们’ve seen with Panda. I 将 suspect that 的re may be other ways 到ccomplish some of those types of targeted goals, so this might 不 necessarily be 的 first iteration of 怎么样 的 algorithm might work. But, I do like seeing an approach spelled out in a way that 将 address some of 的 是sues involving boosting higher 质量 资源s.

  8. 嗨妮维娜娜

    Thank you. 那里 do seem to be some similarities with identifying 参考ring queries –页面似乎基于其中的内容针对特定查询。我们做什么’Google不知道该如何解决该关键字填充域问题。 谷歌确实发布了一项专利,该专利最初是他们于2003年提交的,’授予时间到2011年。我在Google的文章中写过 精确匹配域名专利(检测商业查询), and it presented a number of different ways that 谷歌 might determine whether 要么 不 terms in queries might be commercial, and if so, whether 要么 不 的y should be devalued.

    Doing an analysis to determine whether 要么 不 specific 页s could be associated with 参考ring queries because those terms might be in 的 domain name 是n’等同于停止为同一页面提供查询词的搜索结果,因为查询词在域名中。

  9. 嗨,迈克尔,

    我同意专利的链接分析部分的重要性。我也同意,可能有一项或多项与熊猫程序有关的其他相关专利。我在帖子中提到了基于短语的索引专利,其中有许多专利适用于基于短语的工作方式的不同方面,其中许多申请日期不同。

  10. 嗨亚伦,

    谢谢。 I do like 的 approaches described in 的 patent, and it makes sense to use 的m in a manner such as this. I am expecting to see more on 的 Panda algorithm at some point, and I’m hoping that we don’不必等待太久。

  11. My understanding 是 that a U.S. patent must be filed within 12 months of 的 invention’的首次销售,使用或出版。

    This patent 原为 filed in 2012年9月, more than 18 months after 的 initial Panda update in 2011年2月.

    不会’表示该专利描述的内容不是 熊猫算法?

  12. 嗨,马修,

    如果您访问了针对特定查询进行了优化的Web页面,则很可能会在页面标题,标题,页面内容等中看到关键字词或短语。您可以了解某人可能试图为其优化页面的一个或多个查询。

    PageRank部分基于受欢迎程度,但搜索引擎通常将其称为“quality” measure. It’并非仅基于大量链接,而是基于类似于科学论文中学术引用的概念。重要页面倾向于链接到重要页面的概念是’只需对链接最多的页面进行投票,然后投票’等于《纽约时报》首页上的链接比《福奎尔时报》上的链接重要得多(我的本地报纸每周发表两次)。纽约时报首页上的链接可能比一个小镇报纸的在线首页上的链接成千上万。

    It 是 definitely quite possible that 的re are other related patents out 的re that haven’已公开,但涉及熊猫。

    Thanks regarding 的 new look for 的 现场.

  13. 固这里读比尔!一世’我必须承认,很多事情困扰着我。一世’与y的某个空间相比,这个空间中的一个新手’所有特立独行者!但是,老兄,我非常喜欢阅读和学习您的文章。总是彻底和“complete”! I’与SEO领域中的其他一些人不同,我特别喜欢您对投机的厌恶,但是足够多的是,没有人喜欢那些使用名字命名的人。.继续做这项出色的工作!而且,就像上面的其他人一样,即使我确实错过了“big fish”:-]!很好。现代而干净。真好

    斯科拉。

  14. 感谢您发布此法案,这很有意义。

    Shortly after Panda 原为 rolled out 的re 原为 speculation if 谷歌 是 using a ratio of 的 number of links / number of 页s on a 现场. A number of links / number of 参考ring queries though seems much more interesting.

    你可以这样想– if a 组 of 页s 是 receiving a disproportionately high amount of search 参考rals while having a disproportionate number of links, what does this say about 用户 satisfaction?

    这也可以解释为什么与熊猫打交道的许多传统建议都没有成功–删除重复,空白甚至内容不足的页面都不会’t necessarily reduce 的 number of 参考ring queries. Maybe SEOs shouldn’一直没有删除不’没有内容,而是具有大量查询排名的页面。

  15. Hi 戴夫 ,

    我一触及涉及的专利部分“navigational queries”, that nailed it for me that 的 patent 原为 about 的 Panda update. 那里 原为 some talk from 谷歌 about 的 Panda update being a change 到 document classifier, and 的 patents 谷歌 has published about 航海的 queries are all about classifications of documents.

    的“reference queries” section 是n’关于人们在网站上正在寻找或正在寻找页面的查询,而是页面似乎针对的查询–例如内容服务器场定位的多个关键字变体,例如“How to tie a shoe,” How to knot a shoe,” “如何穿鞋。” That’熊猫,而不是企鹅,而不管专利中的Navneet Panda名称如何。

  16. 嗨,RogueSkolar,

    谢谢。它’很难遗漏一些推测,而且鉴于熊猫和企鹅的更新对许多人及其网站都产生了影响,因此对于每个想法和理论的特定思想和理论肯定会有强烈的情感和情感依恋。更新工作。

    很有可能,如果熊猫和企鹅的更新中可能涉及其他专利,我们将赢得’在授予专利之前,不要看到这些专利,有时专利可能需要一段时间才能通过起诉程序并公开,所以我们不’不知道那里还有什么,什么时候可能成为公共知识。

    Thanks, regarding 的 new look for 的 现场 as well. I’我自己还是很习惯--

  17. 这里’一个有趣的思想实验:暂时忘记此专利拥有Navneet Panda’s name on it. If you were to read it without knowing 的 inventor’的名称,您最想与它关联的是什么Google更新?

    A lot of 的 mechanics described in 的 patent — especially those regarding 的 application of a 组 (site) modification factor to individual search results — fit well with 的 conventional wisdom about 的 Panda algorithm.

    但是,专利中包含的基本价值判断— based on 的 ratio of 独立 links to 参考查询 — doesn’似乎与Google的关系很好’s public statements about Panda, nor with individual stories of Panda 罚款 and recovery. Most of those have centered around content 质量 而不是link factors.

    的patent seems to make 的 assumption that 的re 是 a “natural” ratio between 的 number of 独立 links 到 现场 and 的 number of 参考查询 to that 现场. That 是, for every person who 自然ly creates a link 到ny given 现场, some number of people also perform searches 参考encing that 现场.

    这很直观:人们将链接到一个高质量的网站,他们’重新也要搜索它。没有多少人会链接到低质量的网站。他们也不会搜索它。不论哪个网站’的大小,收到的流量甚至是’s 质量, 的 ratio of links to 参考查询 probably remains fairly consistent.

    谷歌知道这个比率是多少。尽管它可能因利基市场而有所不同(“plurality of 团体 of 资源s”),大多数网站的比例都在Google理解的范围之内。该算法试图惩罚那些超出范围的网站。

    So what sorts of factors might affect 的 ratio of 独立 links to 参考查询?

    Site 质量 seems unlikely 到ffect this ratio. A high-quality 现场 是 going 到ttract both links and 参考查询 (people searching for 的 现场 通过 name). A low-quality 现场 will 不 attract many links, but 是 also unlikely to be searched for 通过 many people. While 的 absolute numbers of links and 参考查询 将 both vary with 质量, 的 ratio between 的 two 是 likely to stay fairly consistent, regardless of 现场 质量.

    什么样的事情 影响链接查询的比例?最明显的一个是建立到低质量站点的不自然链接。让’例如,网站所有者发起了大规模的链接构建活动,并使指向该网站的独立链接数量增加了一倍。这种努力不太可能影响搜索同一网站的人数,因此会导致网站发生重大变化’链接与参考查询的比率,这反过来会影响网站’的组修饰因子。

    What I take away 是 that this specific algorithm 是 better-suited for detecting and penalizing unnatural linking than it 是 for 现场 质量. It 将 also seem to open 的 door wide for negative SEO attacks.

    In those regards, 的 core value judgement embedded in 的 patent seems to be more closely aligned with Penguin than with Panda.

    话虽如此,Navneet Panda’s name 这项专利。我不会’t discount 的 possibility that he 是 responsible for both of those algorithms, nor that this patent describes an amalgamation of 的 two.

  18. 嗨Nemek,

    谢谢。 I do like 的 approach described in 的 patent.

    以某种方式创建的页面,以便它们有目的地响应特定查询,’t anything unusual, but it seems to have been a characteristic of content farm type 现场s that 的 Panda Update seems to be targeting to have 团体 of 页s that 参考ence content in a way that targets lots of specific queries.

    As 的 New York Times 文章, 谷歌’s War on Nonsense 说:

    米勒先生的工作,正如他上周在在线报纸《更快的时报》中的一篇文章中所明确指出的那样,是将有人的研究表明可能需要谷歌提供的字句塞满,将这些字符串定位为标题和标题,并用其他无害的字词修饰它们,使整个甜食模糊地像一篇文章。 AOL会在标题上加上“ Rick Fox胡子”,并押注一定数量的人会将“ Rick Fox胡子”放入Google,并检索Miller先生的文章。

    Getting rid of 的 low 质量 要么 substantially duplicate content 页s probably 原为n’t a bad idea, and 将 be something that I 将 recommend as part of a 现场 audit regardless of Panda. But, it does seem like 的 target should be on 的 content aimed 不 at being 质量 content, but rather aimed more at inducing people to come 到 现场 for specific queries being targeted 通过 页s – 不 参考ring queries, but rather 参考查询.

    再说一次’s 不 unusual for people to create 页s that seem to be about something that people might search for, but 页s that are set up like that, and yet have few 要么 no 独立 links to 的m are probably going to rank lower.

  19. Hi 戴夫 ,

    If you read 的 second part of that first sentence, it says “…and has been classified as 参考ring 到 资源 in 的 组.”

    它没有 ’不要说正在查看的网站页面必须对那些查询词进行排名,尽管可能这样做。它告诉我们Google将查看已提交给它的查询,并查看所讨论网站上的页面是否针对这些查询中的任何一个,或者“refer” to those queries. From 的 passage you quoted, 的 classification of a 参考ral query 是 的 important thing here.

  20. 的patent seems pretty clear that a “reference query” 是 one in which 的 用户 是 trying to reach a specific 现场 要么 页:

    A 参考查询 for a 组 of 资源s 是一个搜索查询 has been submitted 到 search engine and has been classified as 参考ring 到 资源 in 的 组. A query can be classified as 参考ring 到 particular 资源 if 的 query includes a term that 是 recognized 通过 的 system as 参考ring to 的 particular 资源. For example, a term that 参考s 到 资源 may be all of 要么 a portion of a 资源 identifier, e.g., 的 URL, for 的 资源. For example, 的 term “example.com” may be a term that 是 recognized as 参考ring to 的 home 页 of that domain, e.g., 的 资源 whose URL 是 “http://www.example.com”. Thus, search queries including 的 term “example.com” can be classified as 参考ring to that home 页. As another example, if 的 system has data indicating that 的 terms “example sf” and “esf” are commonly used 通过 用户s to 参考 to 的 资源 whose URL 是 “http://www.sf.example.com,” queries that contain 的 terms “example sf” 要么 “esf”, e.g., 的 queries “example sf news” and “esf餐厅点评,” can be counted as 参考查询 for 的 组 that includes 的 资源 whose URL 是 “http://www.sf.example.com.”

    I’m having a hard time squaring that paragraph with your definition and 的 shoe-tying example.

  21. Hi 戴夫 ,

    的query 参考s 到 资源 on 的 现场. So 的 question 是, 怎么样 does that happen?

    那里 are a few different ways. One 将 be that specific 页s on 的 现场 are optimized for those query terms that “refer to 的 资源.”

    Yes, this could be done through on-page optimization, 要么 it could be done 通过 pointing specific anchor text at 页s within 的 资源, 要么 both.

    不幸的是,那没有’告诉我们这是企鹅还是熊猫。

  22. 法案,

    I agree that whether 要么 不 的 现场 ranks for a given query term 是 unmentioned and irrelevant to this section of 的 patent.

    But 我不’t see anywhere in 的 patent that it discusses 的 concept of a 页 参考ring to 要么 targeting a query. (In fact, I see no mention of on-page factors at all.)

    的paragraph I quoted above discusses (and fairly explicitly for a software patent) 的 concept of a 用户 search query that 参考s 到 specific 网站, usually 通过 name 要么 URL:

    A 参考查询 for a 组 of 资源s 是一个搜索查询 has been submitted 到 search engine and has been classified as 参考ring 到 资源 in 的 组.

    Replacing some of 的 cumbersome patent-speak with more common terms makes it even clearer:

    A 参考查询 for a 组 of 资源s 网站 是一个搜索查询 has been submitted 到 search engine and has been classified as 参考ring 参考s资源 in 的 现场.

    的query 参考s to 的 页 要么 现场, 不 的 other way around.

    的examples provided in 的 same paragraph further emphasize that 的y’re talking about a query 参考ring 到 现场, 而不是a 现场 targeting a query:

    search queries including 的 term “example.com” can be classified as 参考ring to that home 页

    queries that contain 的 terms “example sf” 要么 “esf”, e.g., 的 queries “example sf news” and “esf餐厅点评,” can be counted as 参考查询 for 的 组 that includes 的 资源 whose URL 是 “http://www.sf.example.com.”

    I’m 不 trying to pick a fight here, but 的 definition of “reference query” seems pretty central to understanding 的 patent, and 我不’t think that your interpretation 是 supported 通过 的 content of 的 patent.

  23. Hi 戴夫 ,

    I am basing what I am saying on a knowledge of 怎么样 页s rank for something. 的patent’s description provides an example of a 参考查询 that does more than just contain 的 URL 要么 的 name of 的 specific 现场:

    “…queries that contain 的 terms “example sf” 要么 “esf”, e.g., 的 queries “example sf news” and “esf餐厅点评,” can be counted as 参考查询 for 的 组 that includes 的 资源 whose URL 是 “http://www.sf.example.com.”

    That second example of a 参考查询, “esf餐厅点评” doesn’t contain 的 name of 的 URL 要么 的 name of 的 现场.

  24. 嗨托德,

    我认为Google只是在忽略它们时’进行此计算,因此独立链接的数量为’t prejudiced 通过 的 existence of some 现场-wide links. 的idea seems to be to gauge 的 质量 of 页s being examined without 的 influence of dependent links (under 的 same control and/or ownership) and without counting multiple links from 的 same 独立 sources more than once. In other words, 怎么样 likely 是 it that people will link to 页s on 的 现场 (or 团体 of 资源s on 的 现场) without having some kind of relationship with 的 现场? Counting multiple links from 的 same 现场 can throw that off.

  25. Could you please cite 的 portions of 的 patent upon which you’重新得出您的结论?

    我不’t see anything in 的re to suggest that a 参考查询 是 anything other than a query that contains 的 URL 要么 name of a specific 现场.

  26. 只是一个简单的想法

    “这样可以避免站点范围内的链接被多次计数。”

    你觉得呢’s more likely 谷歌 simply ignores 的se links, 要么 that 现场s 的y determine to 不 be 独立 with mass links to each other may actually be caught up in an algo “penalty”?

  27. Hi 戴夫 ,

    ESF绝对是的缩写。“Example SF”, but it’不是域,域和主机名或站点名。为了使Google了解这一点,必须进行某种实际处理。如果我们回到该句子的第一部分并再次查看它,我们会得到有关Google在哪里学到的提示“ESF” 参考s to 的 现场:

    As another example, if 的 system has data indicating that 的 terms “example sf” and “esf” are commonly used 通过 用户s to 参考 to 的 资源 whose URL 是 “http://www.sf.example.com,” queries that contain 的 terms “example sf” 要么 “esf”, e.g., 的 queries “example sf news” and “esf餐厅点评,” can be counted as 参考查询 for 的 组 that includes 的 资源 whose URL 是 “http://www.sf.example.com.

    谷歌可以在几个不同的地方学习可能的内容“commonly used 通过 用户s to 参考 to 的 资源”,并且这些查询可能在查询会话中,其中可能包括对域的搜索,也可能在针对某个网站的一个或多个页面进行了优化的搜索中,例如针对以下查询可能返回“esf餐厅评论。”

    Like 的 参考ring query can also be a 航海的 query, that’s a little like saying that a square can be a rectangle. Not all 参考ring queries are going to be 航海的 queries. In 的 post, I did link 到n earlier post I wrote about 航海的 queries, and a 谷歌 patent for 航海的 queries.

    So, what does any of this have to do with 的 Penguin update? How do you see that fitting in with 的 process described in 的 patent?

  28. 在那个例子中“ESF”的缩写“Example SF” — 的 name of 的 现场 http://www.sf.example.com.

    的next paragraph of 的 patent drives home 的 point, equating 参考查询 with 航海的 queries:

    In addition 要么 in 的 alternative, a query can be categorized as 参考ring 到 particular 资源 when 的 query has been determined to be a 航海的 query to 的 particular 资源. From 的 用户 point of view, a 航海的 query 是 a query that 是 submitted in 要么der to get 到 single, particular web 现场 要么 web 页 of a particular entity.

  29. Hi 戴夫 ,

    搜索引擎不会将“全国足球联赛” 是 nfl.com automatically. It needs to do some kind of analysis, and apply some level of confidence that 的 longer name 是 的 same as nfl. 的patent 参考s to 用户 data as being one source of information about what queries might be classified as 参考ring to 资源 团体 that make up a 现场. That kind of 用户 data can be helpful in making such a determination.

    的patent does 参考 to classifying queries as being 参考ring queries for a 现场. When I stated that a 网站 might be used to help determine what those queries were, I 原为 saying that 的 words on 的 页s of 的 现场 and 的 words that 的 页s of 的 现场 might be optimized for can be determining factors in that classification of queries as 参考ring queries for a 资源.

    至于企鹅,我不知道’看该过程如何用于确定不自然的链接模式’s easy enough just 到nalyze 的 link graph to find things like dense bipartite graphs to find unnatural linking patterns. 的2003 谷歌 patent I mentioned does just that.

    It’这不是一个特定比例的问题或问题,而仅仅是更独立的链接使它更有可能“resource” will be higher 质量 since 的y aren’裙带关系。更多参考查询“resource,”它的可能性越小’s higher 质量. If 的 count 是 close to equal for 的 two (independent links/reference queries), 的n this initial score 是 pretty close to one. This really doesn’不能以任何方式找出不自然的链接模式– it seems instead to be trying to find good links to determine a final score for a 资源.

  30. 出于所有意图和目的“ESF” 是 a name for 的 现场 in question, in 的 same way that “全国足球联赛” and “NFL”都是ww.nfl.com的名称。

    As you correctly point out, 的re’s necessarily a classifier (whose details are 不 described in this patent) that determines whether 要么 不 a term 参考s 到 specific 现场.

    的important point 是 that it’s classifying certain queries as 参考ring to specific web 现场s. It’s classifying web 现场s as 参考ring to specific queries.

    这一切与企鹅有什么关系?不一定。但比较链接到网站的人数(“independent links”)到搜索它的人数(“reference queries”)似乎是检测不自然链接模式的好方法。而且企鹅似乎比熊猫还多。看我的 较早的评论 有关详细信息。

  31. 比尔,我尊重您在这方面的经验,但在这种情况下,我认为’会给您带来损害,并导致您将属性归因于该过程’t supported 通过 的 text of 的 patent.

    Of course 的re’是一个分类器,显然’必须考虑各种因素来确定“全国足球联赛” and “NFL” both 参考 to nfl.com. But for 的 purposes of 的 process described in this patent, 怎么样 的 classifier reaches its conclusions 是 irrelevant. (If 的 充分 classification process 原为 a part of 的 invention, 的 patent 将 have to describe it, which it does 不.)

    根据该专利的定义,“NFL” 是 a 参考查询 to nfl.com, and 只要 to nfl.com. It does 不 参考 到ny other 现场s, regardless of 怎么样 much 的y have tried to optimize for 的 term. A plain reading of 的 patent should make that abundantly clear.

    如果你想灌输“reference query”具有各种各样的特性’专利中提到的,就这样吧。我们可以同意不同意。但是最简单的解释通常是正确的解释,所以我’我会选择接受该术语’实际在专利中定义。

  32. 您好Rajesh

    I’我不确定你会不会“always” viewed 参考查询 的 way that 戴夫 had, since 的re really hasn’t been a 参考ence to 的m like in this patent anywhere else before that I can locate. 的only time you had to begin viewing 的m that way 原为 in reading 的 patent.

    的definition from 的 patent 是 a definition of examples 而不是a definition of what a “reference query”实际上是,但不幸的是’容易忽略一个示例的一部分,该部分明确指出Google会考虑用户数据之类的东西,以等同于“ESF餐厅评论”与示例中的网站。

    If 参考查询 were 只要 queries that included 的 domain name 要么 host name and domain name 要么 现场 name, 的n 的re just 将n’t be many 参考查询 for any one 现场. That 原为 an example, 不 a 充分 definition.

  33. 有趣的帖子,非常详细。一世’m 不 sure, to be honest, 我不’请特别详细地研究Google。

    关于Google的某些事情引起了我的兴趣/对我深深着迷,其中之一是昨天在马特·卡茨(Matt Cutts)身上出现’视频部分。一个人问他“您如何将简单的人气与真正的权威区分开来?”他的回答相当含糊。我记得Panda更新即将推出,人们普遍对这件事感到恐慌。毫无疑问,下一个计划正在进行中,我希望他们将其称为海象。

  34. 法案, a very interesting exchange of thoughts between you and 戴夫 on “reference queries” and I 总是 understood it as 怎么样 戴夫 has done. I truly believe it as 怎么样 用户s use queries to 参考ence 资源s within a 组 and 不 的 otherway round.

    的problem with content farms 原为 的y had links but 的y were unlike wikipedia 要么 NYT as people never used that many 参考查询 to 的m as 的y did while searching for 资源s in wikipedia 要么 NYT. This 是 also confirmed 通过 怎么样 some 现场s like wiki.answers.com still rank very well though 的m seem to have been created like a content farm as you had described above.

    但是有些内容服务器场因其品牌(参考查询)而仍然排名不错。只是搜索“世界上最小的鸟”并查看ask.com,wiki,answers.com在首页上的排名。而且,如果您查看这些页面,它们实际上与您描述内容服务器场的功能相同。但由于其品牌(参考查询)优势,它们的排名仍然不错。

  35. 嗨,亚历克斯,

    谢谢。 I did see 的 video you’re 参考ring to, responding 到 question from AJ科恩 (Blind Five Year Old). 的question 原为:

    “As 谷歌 continues 到dd social signals to 的 algorithm, 怎么样 do you separate simple popularity from true authority?”

    Matt purposefully ignored 的 first part of that question about social signals, but focused upon 的 “简单的人气与真正的权威” part of 的 question.

    I thought his answer 原为 pretty interesting, especially in light of some of 的 discussion above, especially when he starts talking about algorithms working to find evidence that some 现场s might be good matches for specific topical queries.

    我不’不知道我们是否会很快看到另一个大更新,例如熊猫或企鹅,但我们’肯定会看到新的排名信号和方法。海象? --

  36. 法案, I used 的 word “always” to mean 怎么样 i understood 的 term “reference queries” from 的 time i read it on 的 patent.

    以来“Group”该专利中的定义是基于地址的,并且被定义为具有相同的域名或主机名,“reference queries” 将 mean any phrase that includes terms to 参考 到 资源 in 的 组, such terms can 只要 be domain names 要么 host names 要么 anything widely used 通过 searchers to 参考 to that 组. and this could include abbreviations like 美国橄榄球联盟 etc.. So in 的 context of 的 patent, 参考查询 are determined 通过 search 用户s.

    引入熊猫后,另一个有趣的发展是,人们抱怨坐下来引用他们的页面较长的短语或句子进行引用搜索时,将内容排名复制到页面之前(在其网站上)。即使复制站点链接回原始发件人,也据说发生了这种情况(我自己也看到了)。以下内容确实说明了可能发生此问题的线索。

    “As another example, 的 system may have access to data that indicates 怎么样 similar two 资源s are in one 要么 more aspects, e.g., based on whether 的 two 资源s have identical 要么 similar content, identical 要么 similar images, identical 要么 similar formatting, e.g., identical 要么 similar Cascading Style Sheets (CSS), and so on.”

    从上面可以明显看出,复制原始站点内容大部分的站点的链接返回’t being considered as 独立 links and hence 的 value of such link backs have been zeroed out for 的 target 现场 通过 的 “independence”熊猫算法中传入链接的因子。

  37. 首先,感谢比尔提请我们注意该专利。一世’自2011年4月以来一直在研究熊猫并试图弄清楚–所以这是一个巨大的发现。

    Having analysed 的 patent in detail –我必须承认我的理解是 “reference queries”几乎是“navigational queries” –即。用于查找特定网站的查询。

    因此,当您说(上述)时:

    ———————————-
    的more 参考查询 for a “resource,”它的可能性越小’s higher 质量. If 的 count 是 close to equal for 的 two (independent links/reference queries), 的n this initial score 是 pretty close to one.
    ———————————-

    我很惊讶– because my understanding 原为 的 exact opposite: For a given number of backlinks, 的 more 参考查询 a 现场 has, 的 better 质量 (or actually 的 bigger 牌) it’s likely to be.

    所以举个例子–搜索者更可能键入:

    “纽约时报马来西亚航空航班”
    要么
    “美国有线电视新闻网(CNN)巴拉克·奥巴马(Barak Obama)讲话”
    要么
    “WebMD Blood Pressure”

    而不是“Noname垃圾网站血压”

    然后’s because 的 big 牌 现场s have built a reputation for 质量 that searchers recognise –他们在想要重复的网站上拥有良好的体验– and hence search for 的m 通过 name. Hence 的 correlation 谷歌 has found between 航海的 searches and higher 质量.

    记得马特·卡茨在这里说过: http://www.wired.com/2011/03/the-panda-that-hates-farms/2/

    ——–
    And we actually came up with a classifier to say, okay, IRS 要么 Wikipedia 要么 New York Times 是 over on this side, and 的 low-quality 现场s are over on this side. And you can really see mathematical reasons
    ——–

    So he specifically mentions big 牌 现场s that people are more likely to do 航海的 searches for.

    这就是我们在熊猫之后看到的– a huge boost for 的 major 牌s.

    对于所有熊猫’s shortcomings –这是区分人们会积极寻找的站点的一种非常聪明的方法–以及那些只擅长SEO的人。

    I’我也在自己的市场上做了一些研究– and saw some indication that people who did well after Panda have a better ratio of 航海的 searches to links than those who did badly – (but it’是一个非常小的样本,当然不是结论性的)。

    My plan for my Panda-hit 现场 having studied 的 patent 是 to try to get more 自然 航海的 searches 通过 boosting our 牌 recognition. But you’我读了很多专利,所以我希望’m 不 heading in precisely 的 wrong direction…

  38. 嗨,比尔,
    我真的不知道’t see 怎么样 this could possibly be 的 initial Panda patent. Panda 原为 very clearly about 谷歌 making a judgement call on 现场 质量, and 我不’看不到该专利中描述的内容足以证明其合理性‘the’ Panda patent.

    Yes, I think this 是 probably part of 的 Panda algorithm as it currently stands, but 我不’t agree that this describes 的 initial update.

    你觉得呢 是 possible that this 原为 part of an update to 的 充分 Panda algorithm (e.g. http://www.seroundtable.com/google-panda-20-15789.html)?

  39. 嗨,帕特里克,

    该专利从已经根据相关性和PageRank之类的查询对网页进行排名的页面开始。

    It 是 完成ly and totally about 现场 质量 and about 页 classification. Given that, I think it could easily describe 的 initial update. 我不’我不知道搜索引擎监视或搜索引擎圆桌会议对Panda Updates进行编号的原因是什么,但我想它们的数量对Google几乎没有意义。我不会在该帖子中称搜索引擎圆桌会议链接到的Panda版本“full” Panda algorithm.

  40. 感谢您发表另一篇深入的文章-(每次都会使我的头旋转,但是’s 总是 a welcomed one)- I’虽然这与它在栅栏上’s leaning slightly more to this being actually Panda 的n 不.

  41. 法案,
    再一次,你’ve shed light to 的 likes of me who take 谷歌’的专利是象形文字。

    I’d like to cite this as a 资源 and post this a step simpler in my blog. This has definitely shed some light to 共引. That 的re are more specific ways to do it than just having your 牌 cited in a 网站.

  42. 嗨,肖恩,

    This has 不hing to do with 共引, so I’d提醒您这样做。它也没有’t mention 的 word “brand” at all, and I’我已经看到了很多关于此内容的博客文章,这些文章过分夸大了。

  43. 嗯,

    所以等等,我是否理解这句话是错误的:
    “It might also count 默示 links, which sound more like what we often tend to 参考 到s citations. An express link can be used to navigate 到 place, where an 默示 link can’点击以将一个人带到该链接的目标。”

    Because 怎么样 I understand it 是 that if you 共同引用 in such a way that it 是 a 航海的 indication, it affects rankings somewhat.

    如果我纠正我’m wrong.

  44. 嗨肖恩

    我不’无法理解您在写时要说的话:“Because 怎么样 I understand it 是 that if you 共同引用 in such a way that it 是 a 航海的 indication, it affects rankings somewhat.”

    我也会认真建议不要使用“co-cite” 要么 “co-citation,”特别是如果您的意思是某些东西可能是在moz或seomoz上写过的,因为有些关于共现的糊涂文章使用了该术语,“co-citation”相反地​​。

  45. 法案,

    That clarifies things. Knowing this, I definitely have to look deeper into 的 differences between “Co-citation” and “Co-occurrence” so 我不’t make 的 mistake of using one when I mean 的 other.

    您是否愿意帮助我理解此词在外行中的含义’s terms:
    “It might also count 默示 links, which sound more like what we often tend to 参考 到s citations. An express link can be used to navigate 到 place, where an 默示 link can’点击以将一个人带到该链接的目标。”

    谢谢比尔!

  46. 嗨肖恩

    An “implied”页面上的链接是创建页面的人没有’t插入到该页面的实际链接,但可能包含该页面的URL(以不可点击的方式)或引用了该URL。“the homepage of ESPN”或以不’t可点击的。一个例子是引文(引文,‘co-citation”),就像我们在本地搜索中经常提到的那种一样,在这里我们可以看到商家的名称以及一些其他地理信息,例如电话号码或街道地址的一部分或全部。

    那里 have been a couple of 文章s posted which have made a big deal of express links verses 默示 links, and some kind of ratio between 的 two (which 是 不 part of this patent at all, but if you’re 不 careful when reading it might think you see). 那里 是 a 参考ence 到 ratio, but it’s a ratio of “independent links” to “referring queries.”

    It’更好地理解涉及到该专利本身的同被引与同时存在之间的区别和区别,确实是在浪费您的时间,因为该专利与二者均无关。如果您要研究该主题,请忽略在Moz上有关同被引用或同时出现的任何内容,因为Rand在他撰写有关该主题的最后几次中才混淆了很多事情(我讨厌不得不写那,但是’是的)。希望兰德能再试一次,并弄清楚(我希望)。

  47. 法案,

    我懂了。这就解释了。
    Knowing this, has 的re ever been a patent that tells us that 共引 and co-occurrence 是 a factor when determining rankings? Or 是 it just a baseless hype that 的 big blogs have predicted 要么, worse, concocted?

    非常感谢您抽出宝贵的时间来澄清事情。

  48. 嗨,肖恩,

    那里 are a handful of patents from 谷歌 that mention 共引, but 的y primarily focus upon 怎么样 谷歌 might use 共引 to find 现场s that are similar, like when you see in search results a message that similar 现场s to one you are looking up might be example.com, example1.com, example2.com, etc. This 是 because those 现场s tend to be linked to 通过 many of 的 same 现场s (or “co-cited”).

    有超过100项已授予Google的已获授权专利,其中包括并讨论共现,其中涉及查询重写和其他主题,其中可能包括针对特定查询在一组SERPS中提高某些结果。一世’ve在其中一些上写了博客文章(我还有很多其他文章’也写过关于共现的文章:

    谷歌如何根据查询会话中的同时出现来改革查询
    基于单词之间的关系的排名网页(GOOGLE的同频专利)
    谷歌如何用同频替换查询条款

  49. Great 文章 法案, I am a bit of a newb when it comes to understanding 的 way 谷歌 works but 文章s like this really help gain a better understanding of this update.

  50. I’m a bit late to 的 party here and I’m 不 even clear why this 是 even a question because it 原为 充分y and contemporaneously answered 通过 Matt and Amit in 的 interview here: //www.wired.com/2011/03/the-panda-that-hates-farms/

    For us geeks, 的 algorithm 是 pretty clearly explicated as a feature based 机器学习 algorithm that uses (chiefly) link text, link topology, and agreement with 现场 content as 的 prime features.

  51. I am pretty certain that this 是 的 Panda patent. I 原为 working with a network of 现场s that had most of 的 现场s slammed 通过 Panda 1.0. I 原为 总是 baffled as to why a few of 的 现场s, with just as much problematic content, were untouched and even benefited from 的 update.

    的fact that network links were taken into account in 的 algo update now makes perfect sense to me. 的entire network crosslinked 的 hell out of each other, but 的 handful of 现场s that survived 的 update actually had a significant amount of good non-network links.

  52. Thanks a lot . I 原为 不 clear on Panda at all reading so many 文章s around 的 web. but this 是 really good. Thanks

  53. Late to 的 party, as usual.

    阅读了所有讨论之后, ’s tough to square 的 patent interpretation presented here. 的confusion centers around definition of “reference query”. Bill points out “reference query” hasn’以前没有使用过,因此似乎是了解该专利内容的关键。

    从评论者开始-“Dave”(和其他借调他的人)找到“reference query”品牌相关搜索的同义词。 (以及导航查询。)还有一个“natural” relationship between volume of links and search volume for particular 牌 (or 牌 related terms). A high # links 只要 自然 if 牌 related search volume 是 also high. Site with no 牌 recognition 将 be penalized as unnatural if it had a high number of links (which pushed it to 的 top), but paltry volume of 牌 searching 要么 awareness.

    的depressing aspect of this interpretation 是 one of 的 strong getting stronger. Strong 牌s are allowed high number of links and thus high search visibility. Seems like a horrible search engine to me.

    相比之下,比尔读“reference query”在更广泛,更具弹性的环境中。它’页面要定位的术语范围。广泛定位与狭窄定位之间存在反比关系“independent” link base signifies low 质量. (That ratio less than 1). 的examples provided in patent using 牌 related queries obscure as examples because 的y do indeed suggest 牌 specificity, but need 不 只要 be about 牌 specificity.

    任何人,莱斯利·罗德(Leslie Rohde),我的optilink赢了’t update, you’重新杀死我-小

    大家好

评论被关闭。