作者: Haralambos Marmanis / Dmitry Babenko
出版社: Manning Publications
出版年: April 28, 2009
页数: 368
定价: USD 44.99
装帧: Paperback
ISBN: 9781933988665
出版社: Manning Publications
出版年: April 28, 2009
页数: 368
定价: USD 44.99
装帧: Paperback
ISBN: 9781933988665
内容简介 · · · · · ·
Web 2.0 applications provide a rich user experience, but the parts you can't see are just as important-and impressive. They use powerful techniques to process information intelligently and offer features based on patterns and relationships in data. Algorithms of the Intelligent Web shows readers how to use the same techniques employed by household names like Google Ad Sense, Ne... (展开全部)
Web 2.0 applications provide a rich user experience, but the parts you can't see are just as important-and impressive. They use powerful techniques to process information intelligently and offer features based on patterns and relationships in data. Algorithms of the Intelligent Web shows readers how to use the same techniques employed by household names like Google Ad Sense, Netflix, and Amazon to transform raw data into actionable information.
Algorithms of the Intelligent Web is an example-driven blueprint for creating applications that collect, analyze, and act on the massive quantities of data users leave in their wake as they use the web. Readers learn to build Netflix-style recommendation engines, and how to apply the same techniques to social-networking sites. See how click-trace analysis can result in smarter ad rotations. All the examples are designed both to be reused and to illustrate a general technique- an algorithm-that applies to a broad range of scenarios.
As they work through the book's many examples, readers learn about recommendation systems, search and ranking, automatic grouping of similar objects, classification of objects, forecasting models, and autonomous agents. They also become familiar with a large number of open-source libraries and SDKs, and freely available APIs from the hottest sites on the internet, such as Facebook, Google, eBay, and Yahoo.
Algorithms of the Intelligent Web is an example-driven blueprint for creating applications that collect, analyze, and act on the massive quantities of data users leave in their wake as they use the web. Readers learn to build Netflix-style recommendation engines, and how to apply the same techniques to social-networking sites. See how click-trace analysis can result in smarter ad rotations. All the examples are designed both to be reused and to illustrate a general technique- an algorithm-that applies to a broad range of scenarios.
As they work through the book's many examples, readers learn about recommendation systems, search and ranking, automatic grouping of similar objects, classification of objects, forecasting models, and autonomous agents. They also become familiar with a large number of open-source libraries and SDKs, and freely available APIs from the hottest sites on the internet, such as Facebook, Google, eBay, and Yahoo.
作者简介 · · · · · ·
Dr. Haralambos (Babis) Marmanis is a pioneer in the adoption of machine learning techniques for industrial solutions, and also a world expert in supply management. He has about twenty years of experience in developing professional software. Currently, he is the director of R&D and chief architect, for expense management solutions, at Emptoris, Inc. Babis holds a Ph.D. in applie... (展开全部)
Dr. Haralambos (Babis) Marmanis is a pioneer in the adoption of machine learning techniques for industrial solutions, and also a world expert in supply management. He has about twenty years of experience in developing professional software. Currently, he is the director of R&D and chief architect, for expense management solutions, at Emptoris, Inc. Babis holds a Ph.D. in applied mathematics from Brown University, an M.S. degree in theoretical and applied mechanics from the University of Illinois at Urbana-Champaign, and B.S. and M.S. degrees in civil engineering from the Aristotle University of Thessaloniki in Greece. He was the recipient of the Sigma Xi award for innovative research in 2000, and he is the author of numerous publications in peer-reviewed international scientific journals, conferences, and technical periodicals.
Dmitry Babenko is the lead for the data warehouse infrastructure at Emptoris, Inc. He is a software engineer and architect with 13 years of experience in the IT industry. He has designed and built a wide variety of applications and infrastructure frameworks for banking, insurance, supply-chain management, and business intelligence companies. He received a M.S. degree in computer science from Belarussian State University of Informatics and Radioelectronics.
Dmitry Babenko is the lead for the data warehouse infrastructure at Emptoris, Inc. He is a software engineer and architect with 13 years of experience in the IT industry. He has designed and built a wide variety of applications and infrastructure frameworks for banking, insurance, supply-chain management, and business intelligence companies. He received a M.S. degree in computer science from Belarussian State University of Informatics and Radioelectronics.
豆瓣成员常用的标签(共31个) · · · · · ·
喜欢读"Algorithms of the Intelligent Web"的人也喜欢 · · · · · ·
按有用程度 按页码先后 最新笔记
-
第84页
《Algorithms of the Intelligen Web》这本书是09年时一位老师推荐给我的,当时我想用Lucene做点东西,那位老师就推荐我去看这本书的第二章,当时我有很多打印纸无处可用,就把这本书打印出来看了(这样好像也违反了许可吧。。。),看完Chapter 2就顺带看了 Chapter 3 Creating suggestions and recommendations。去年10月份在豆瓣上看到有人在翻译这本书,想起来了就翻出来复习复习,也把其他几章看下。看到84页页首附近的这句话... (更多)《Algorithms of the Intelligen Web》这本书是09年时一位老师推荐给我的,当时我想用Lucene做点东西,那位老师就推荐我去看这本书的第二章,当时我有很多打印纸无处可用,就把这本书打印出来看了(这样好像也违反了许可吧。。。),看完Chapter 2就顺带看了 Chapter 3 Creating suggestions and recommendations。去年10月份在豆瓣上看到有人在翻译这本书,想起来了就翻出来复习复习,也把其他几章看下。看到84页页首附近的这句话
我09年时在这句话旁边做了红色的笔记,大意是认为users或items间的相似度不一定要是对称的。去年下半年以来看过一些推荐系统相关的论文,涉及到的相似度概念都是对称的,我也闪现过几次认为可以非对称的念头。但没有深入思考,现在在这里把想法整理一下。推荐系统中similarity这个概念的直接意义当然是描述users或items间的相似的程度,若仅从这个直接的意义上看相似度应该是对称的,毕竟“相似度”都有个“相”字在里面。但similarity的意义并不仅仅停留于此,其终极意义应该是为产生更好的recommendations而提供服务,similarity对称与否在这个最终的意义下是都是浮云。以user A和user B为例,把本质抽出来,在追求为A提供更好的推荐这个目标的过程中,A与B以及其他users的similarity起到的本质作用就是【刻画B及其他users与A的关系对于为A产生更好推荐的作用(其实就是权重)】。从这个观点出发的话,A与B的similarity就不应是对称的了,因为这种作用不应当是对等的。在生活中有很多例子,比如专业学习方面,老师对items的ratings信息对为学生提供推荐的作用就比反过来大。下面简单地提供个不对称相似度的度量模型。以user A与user B为例,设A,B评过分的items集合分别为items(A),items(B)。考虑两个特殊情况:1) items(A)与items(B)没有交集可以认为A与B是兴趣爱好不同的人,也即对A来说与对B来说相似度均为02)items(A)与items(B)相等 可以认为A与B是兴趣爱好相同的人,也即对A来说与对B来说的相似度均为1那么对一般的情况就可以定义对A而言B与A的相似度Similarity(A,B) = |intersection(A,B)|/|items(A)|.相应地,对B而言A与B的相似度Similarity(B,A)=|intersection(A,B)|/|items(A)|上面的不对称相似度,没有考虑A,B对intersection(A,B)中items评分值的差异带来的唱响,我觉得可以将上面的不对称相似度作为一个系数,乘上常用的对称相似度系数。ACM的新论文《The Wisdom of the Few》认为基于专家的协同过滤能产生更好的效果,我觉得不对称的相似度模型与这篇论文的思想有相通之处(虽然这篇文章中使用的也是对称的相似度),就是把users区分开,不要看成同一的。同样的信息内容,对于不同的人而言信息量是不同的,这也符合信息论。思考的还不成熟,有时间应该用MovieLen的数据与常用的对称相关度模型做做对比实验,希望春节期间有时间做。见笑于方家了。 (收起)the similarity matrix is symmetrical. This simply means that if user A is similarity to user B with a similarity value X then user B will be similar to user A with a similarity value equal to X.
2011-01-27 00:26:54 5回应
-
第118页
Bell and Keorren are leading the Netflix prize competitioin (at the time of this writing), and their assessment was the following: "we found no perfect models. Instead, our best results came from combining predictions of models that complemented each other." 这就像艾滋病的鸡尾酒疗法? (更多)
这就像艾滋病的鸡尾酒疗法? (收起)Bell and Keorren are leading the Netflix prize competitioin (at the time of this writing), and their assessment was the following: "we found no perfect models. Instead, our best results came from combining predictions of models that complemented each other."
2011-01-28 09:49:48 3回应
-
第84页
《Algorithms of the Intelligen Web》这本书是09年时一位老师推荐给我的,当时我想用Lucene做点东西,那位老师就推荐我去看这本书的第二章,当时我有很多打印纸无处可用,就把这本书打印出来看了(这样好像也违反了许可吧。。。),看完Chapter 2就顺带看了 Chapter 3 Creating suggestions and recommendations。去年10月份在豆瓣上看到有人在翻译这本书,想起来了就翻出来复习复习,也把其他几章看下。看到84页页首附近的这句话... (更多)《Algorithms of the Intelligen Web》这本书是09年时一位老师推荐给我的,当时我想用Lucene做点东西,那位老师就推荐我去看这本书的第二章,当时我有很多打印纸无处可用,就把这本书打印出来看了(这样好像也违反了许可吧。。。),看完Chapter 2就顺带看了 Chapter 3 Creating suggestions and recommendations。去年10月份在豆瓣上看到有人在翻译这本书,想起来了就翻出来复习复习,也把其他几章看下。看到84页页首附近的这句话
我09年时在这句话旁边做了红色的笔记,大意是认为users或items间的相似度不一定要是对称的。去年下半年以来看过一些推荐系统相关的论文,涉及到的相似度概念都是对称的,我也闪现过几次认为可以非对称的念头。但没有深入思考,现在在这里把想法整理一下。推荐系统中similarity这个概念的直接意义当然是描述users或items间的相似的程度,若仅从这个直接的意义上看相似度应该是对称的,毕竟“相似度”都有个“相”字在里面。但similarity的意义并不仅仅停留于此,其终极意义应该是为产生更好的recommendations而提供服务,similarity对称与否在这个最终的意义下是都是浮云。以user A和user B为例,把本质抽出来,在追求为A提供更好的推荐这个目标的过程中,A与B以及其他users的similarity起到的本质作用就是【刻画B及其他users与A的关系对于为A产生更好推荐的作用(其实就是权重)】。从这个观点出发的话,A与B的similarity就不应是对称的了,因为这种作用不应当是对等的。在生活中有很多例子,比如专业学习方面,老师对items的ratings信息对为学生提供推荐的作用就比反过来大。下面简单地提供个不对称相似度的度量模型。以user A与user B为例,设A,B评过分的items集合分别为items(A),items(B)。考虑两个特殊情况:1) items(A)与items(B)没有交集可以认为A与B是兴趣爱好不同的人,也即对A来说与对B来说相似度均为02)items(A)与items(B)相等 可以认为A与B是兴趣爱好相同的人,也即对A来说与对B来说的相似度均为1那么对一般的情况就可以定义对A而言B与A的相似度Similarity(A,B) = |intersection(A,B)|/|items(A)|.相应地,对B而言A与B的相似度Similarity(B,A)=|intersection(A,B)|/|items(A)|上面的不对称相似度,没有考虑A,B对intersection(A,B)中items评分值的差异带来的唱响,我觉得可以将上面的不对称相似度作为一个系数,乘上常用的对称相似度系数。ACM的新论文《The Wisdom of the Few》认为基于专家的协同过滤能产生更好的效果,我觉得不对称的相似度模型与这篇论文的思想有相通之处(虽然这篇文章中使用的也是对称的相似度),就是把users区分开,不要看成同一的。同样的信息内容,对于不同的人而言信息量是不同的,这也符合信息论。思考的还不成熟,有时间应该用MovieLen的数据与常用的对称相关度模型做做对比实验,希望春节期间有时间做。见笑于方家了。 (收起)the similarity matrix is symmetrical. This simply means that if user A is similarity to user B with a similarity value X then user B will be similar to user A with a similarity value equal to X.
2011-01-27 00:26:54 5回应
-
第118页
Bell and Keorren are leading the Netflix prize competitioin (at the time of this writing), and their assessment was the following: "we found no perfect models. Instead, our best results came from combining predictions of models that complemented each other." 这就像艾滋病的鸡尾酒疗法? (更多)
这就像艾滋病的鸡尾酒疗法? (收起)Bell and Keorren are leading the Netflix prize competitioin (at the time of this writing), and their assessment was the following: "we found no perfect models. Instead, our best results came from combining predictions of models that complemented each other."
2011-01-28 09:49:48 3回应
-
第118页
Bell and Keorren are leading the Netflix prize competitioin (at the time of this writing), and their assessment was the following: "we found no perfect models. Instead, our best results came from combining predictions of models that complemented each other." 这就像艾滋病的鸡尾酒疗法? (更多)
这就像艾滋病的鸡尾酒疗法? (收起)Bell and Keorren are leading the Netflix prize competitioin (at the time of this writing), and their assessment was the following: "we found no perfect models. Instead, our best results came from combining predictions of models that complemented each other."
2011-01-28 09:49:48 3回应
-
第84页
《Algorithms of the Intelligen Web》这本书是09年时一位老师推荐给我的,当时我想用Lucene做点东西,那位老师就推荐我去看这本书的第二章,当时我有很多打印纸无处可用,就把这本书打印出来看了(这样好像也违反了许可吧。。。),看完Chapter 2就顺带看了 Chapter 3 Creating suggestions and recommendations。去年10月份在豆瓣上看到有人在翻译这本书,想起来了就翻出来复习复习,也把其他几章看下。看到84页页首附近的这句话... (更多)《Algorithms of the Intelligen Web》这本书是09年时一位老师推荐给我的,当时我想用Lucene做点东西,那位老师就推荐我去看这本书的第二章,当时我有很多打印纸无处可用,就把这本书打印出来看了(这样好像也违反了许可吧。。。),看完Chapter 2就顺带看了 Chapter 3 Creating suggestions and recommendations。去年10月份在豆瓣上看到有人在翻译这本书,想起来了就翻出来复习复习,也把其他几章看下。看到84页页首附近的这句话
我09年时在这句话旁边做了红色的笔记,大意是认为users或items间的相似度不一定要是对称的。去年下半年以来看过一些推荐系统相关的论文,涉及到的相似度概念都是对称的,我也闪现过几次认为可以非对称的念头。但没有深入思考,现在在这里把想法整理一下。推荐系统中similarity这个概念的直接意义当然是描述users或items间的相似的程度,若仅从这个直接的意义上看相似度应该是对称的,毕竟“相似度”都有个“相”字在里面。但similarity的意义并不仅仅停留于此,其终极意义应该是为产生更好的recommendations而提供服务,similarity对称与否在这个最终的意义下是都是浮云。以user A和user B为例,把本质抽出来,在追求为A提供更好的推荐这个目标的过程中,A与B以及其他users的similarity起到的本质作用就是【刻画B及其他users与A的关系对于为A产生更好推荐的作用(其实就是权重)】。从这个观点出发的话,A与B的similarity就不应是对称的了,因为这种作用不应当是对等的。在生活中有很多例子,比如专业学习方面,老师对items的ratings信息对为学生提供推荐的作用就比反过来大。下面简单地提供个不对称相似度的度量模型。以user A与user B为例,设A,B评过分的items集合分别为items(A),items(B)。考虑两个特殊情况:1) items(A)与items(B)没有交集可以认为A与B是兴趣爱好不同的人,也即对A来说与对B来说相似度均为02)items(A)与items(B)相等 可以认为A与B是兴趣爱好相同的人,也即对A来说与对B来说的相似度均为1那么对一般的情况就可以定义对A而言B与A的相似度Similarity(A,B) = |intersection(A,B)|/|items(A)|.相应地,对B而言A与B的相似度Similarity(B,A)=|intersection(A,B)|/|items(A)|上面的不对称相似度,没有考虑A,B对intersection(A,B)中items评分值的差异带来的唱响,我觉得可以将上面的不对称相似度作为一个系数,乘上常用的对称相似度系数。ACM的新论文《The Wisdom of the Few》认为基于专家的协同过滤能产生更好的效果,我觉得不对称的相似度模型与这篇论文的思想有相通之处(虽然这篇文章中使用的也是对称的相似度),就是把users区分开,不要看成同一的。同样的信息内容,对于不同的人而言信息量是不同的,这也符合信息论。思考的还不成熟,有时间应该用MovieLen的数据与常用的对称相关度模型做做对比实验,希望春节期间有时间做。见笑于方家了。 (收起)the similarity matrix is symmetrical. This simply means that if user A is similarity to user B with a similarity value X then user B will be similar to user A with a similarity value equal to X.
2011-01-27 00:26:54 5回应
书评 · · · · · · 我来评论这本书
热门评论 最新评论
第一个在"Algorithms of the Intelligent Web"的论坛里发言
在哪儿买这本书? · · · · · ·
- 亚马逊 (RMB 268.20)
- 查看1家网店价格 (268.20元起)
- 加入购书单 多本比价 批量购买 已在购书单
以下豆列推荐 · · · · · · (全部)
- 数据挖掘与个性化推荐 (刘未鹏(pongba))
- 机器学习应用 (Fan)
- 编程书籍候选 (笨蛋.mk⑨)
- !~search engine & data mining (∫跬步→千里)
- 推荐系统 (sesamfox)
谁读这本书?
喜欢这本书的人常去的小组 · · · · · ·

- 分布式技术与系统软件 (697)

- 个性化推荐 (404)

- 华章书友俱乐部 (270)

- drupal (398)

- WebRebuild (458)

- 生物信息学 Bioinformatic... (2284)

- git (781)

- 博文视点交流组 (419)
喜欢这本书的人关注的活动 · · · · · ·
订阅关于Algorithms of the Intelligent Web的评论:
feed: rss 2.0











