The Long Tail - 1st part

本文探讨了亚马逊如何通过其推荐系统让一本原本默默无闻的书籍《Touching the Void》成为畅销书。通过分析购买行为模式并推荐类似产品,亚马逊不仅提升了这本书的销量,还揭示了消费者偏好可能远比市场主流更为多样。
Orginal:
http://www.wired.com/wired/archive/12.10/tail.html
[color=darkred]I type them, not copy, just for English learning.[/color] :D

Chinese: http://yeeyan.com/articles/view/thunder/23

In 1988, a British mountain climber named Joe Simpson wrote a book called Touching the Void, a harrowing account of near death in the Peruvian Andes. It got good reviews but, only a modest success, it was soon forgotten. Then, a decade later, a stranger thing happened. Jon krakauer wrote a Into Thin Air, another book about a mountain-climing tragedy, which became a publishing sensation. Suddenly Touching the Void stared to sell again.

Random House rushed out a new edition to keep up with demand. Booksellers began to promote it next their Into thin Air displays, and sales rose further. A revised paperback edition, which came out in January, spent 14 weeks on the New York Times bestseller list. That same month, IFC Films released a docudrama of the story to critical acclaim. Now Touching the Void outsells Into Thin Air more than two or one.

What happened? In short, Amazon.com recommendations. The online bookseller's software noted patterns in buying behavior and suggested that reader who liked Into Thin Air would also like Touching the Void. People took the suggestion, agreed wholeheartedly, wrote rhapsodic reviews. More sales, more algorithm-fueled recommendations, and the positive feedback loop kicked in.

Particularly notable is that when Krakauer's book hit shelves, Simpon's was nearly out of print. A few years ago, readers of Krakauer would never even have learned about Simpson's book- and if they had, they wouldn't have been able to find it. Amazon changed that. It created the Touching the Void phenomenon by combining infinite shelf space with real-time information about buying trends and public opinion. The result: rising demand for an obscure book.

This is not just a virtue of online booksellers; it is an example of an entirely new economic model for th media and entertainment industries, one that is just beginning to show its power. Unlimitd selection is revealing truths about what consumers want and how they want ti get it in service after service, from DVDs at Netflix to music videos on Yahoo! Launch to songs in the iTunes music store and Rhapsody. People are going deep into the catalog, down the long, long list of available titles, far past what's available at Blockbuster Video, Tower Record, and Barnes & Noble. And the more they found, the more they like. As they wander further from the beaten path, they discover their tastes is not as mainstream as they thought (or as they had been led to believe by marketing, a lack of alternatives, and a hit-driven culture).

An analysis of the sales data and trends from services and others like them shows that emerging digital entertainment economy is going to be radically different from today's mass market. If the 20th-century was about hits, the 21st will be equally about misses.

For too long we've been suffering the tyranny of lowest-common-denominator fare, subjected to brain-dead summer blockbusters and manufactured pop. Why? Economics. Many of our assumptions about popular taste are actually artifacts of poor supply-and-demand matching - a market response to inefficient distribution.

The main problem, if that's the word, is that we liev in the physical world and, until recently, most of our entertainment media did, too. But that world puts two dramatic limitations on our entertainment.

The first is the need to find local audiences. An average movie theater will not show a film unless it can attract at least 1,500 people over a two-week run; that's essentially the rent for a screen. An average record store needs to sell at least two copies of a CD per year to make it worth carrying; that's the rent for a half inch of shelf space. And so on for DVD retal shops, videogame stores, booksellers, and newsstands.

In each case, retailers will carry only content that can generate sufficient demand to earn its keep. But each can pull only from a limited local population - perhaps a 10-mile radius for a typical movie theater, less than that for music and bookstores, and even less (just a mile or two) for video rental shops. It's not enough for a great documentary to have a potential national audience of half a million;; what matters is how many it has in the northern part of Rockville, Maryland, and among the mall shoppers of Walnut Creek, California.

There is plenty of great entertainment with potentially large, even rapturous, national audiences that can not clear that bar. For instance, The Triplets of Belleville, a critically acclaimed film that was nominated for the best animated feature Oscar this year, opened on just six screens nationwide. An even more striking example is the plight of Bollywood in America. Each year, India's film industry puts out more than 800 feature films. There are an estimated 1.7 million Indians in the US. Yet the top-rated (according to Amazon's Internet Movie Database) Hindi-language film, Lagaan: Once Upon a Time in India, opened on just two screens, and it was one of only a handful of Indian films to get any US distribution at all. In the tyranny of physical space, an audience too thinly spread is the same as no audience at all.

The othe constraint of the physical world is physics itself. The radio spectrum can carry only so many stations, and a coaxial cable so many TV channels. And, of course, there are only 24 hours a day of programming. The curse of broadcast technologies is that they are profligate users of limited resources. The result is yet another instance of having to aggregate large audiences in one geographic area - another high bar, above which only a fraction of potential content rises.

The past century of entertainment has offered an easy solution to these constraints. Hits fill theaters, fly off shelved, and keep listeners and viewers from touching their dials and remotes. Nothing wrong with that; indeed, sociologists will tell you that hits are hardwired into human psychology, the combinatorial effect of conformity and word of mouth. And to be sure, a healthy share of hits earn their place: Great songs, movies, and books attract bog, broad audiences.
帮我对着正确代码调试错误代码并不要改变错误代码的变量名: 错误代码: #include<bits/stdc++.h> using namespace std; const int MAXN = 3005; int N, K; int a[MAXN][MAXN], st[MAXN][MAXN][3], q[MAXN]; //st[i][j][k]表示从(i,j)开始,大小(边长)为2^k的三角形内最大值,第三维用滚动数组优化 int main(){ cin >> N >> K; for( int i = 1; i <= N; i ++ ){ for( int j = 1; j <= i; j ++ ){ cin >> a[i][j]; st[i][j][0] = a[i][j]; } } for( int t = 1; t <= log2(K); t ++ ){ for( int i = 1; i <= N; i ++ ){ int head = 1, tail = 0; int len = 1 << t;//记录窗口大小 int x = i + ( 1 << t - 1 );//三角形底的行 for( int j = 1; j <= N; j ++ ){//当前窗口:[j−len+1, j] q[head] 表示当前窗口最左侧的列编号 st[i][j][t % 2] = st[i][j][( t - 1 ) % 2];//类似初始化 while( head <= tail && q[head] <= j - len ) head ++;//去头 while( head <= tail && st[x][j][( t - 1 ) % 2] >= st[x][q[tail]][t % 2] ) tail --;//去尾 /* 对于去尾操作的一些解释:(from ai) 新元素st[x][j][(l−1)%2] 表示当前行x的列j的值。 若st[x][j]>st[x][q[tail]],说明队尾元素q[tail] 的值小于新元素,且未来不可能成为最大值(因j更靠后,存活时间更长)。 */ q[++ tail] = j;//更新 if( j >= len )//取出最大值 st[i][j - len + 1][t % 2] = max( st[i][j - len + 1][t % 2], st[x][q[head]][(t - 1) % 2 ] ); } } } long long ans = 0; for( int i = 1; i + K - 1 <= N; i ++ ){//枚举子三角形矩阵起始行(顶点) for( int j = 1; j <= i; j ++ ){//枚举起始列(为三角形区域) int len = log2( K ); int maxx = st[i][j][len % 2]; //将k × k拆分成两个重叠区域 int x = i + K - ( 1 << len ); for( int y = j; y <= j + K - ( 1 << len ); y ++ )//又是滑动窗口 maxx = max( maxx, st[x][y][len % 2] ); ans += maxx; } } cout << ans; return 0; } 正确代码: #include <bits/stdc++.h> using namespace std; const int N = 3005; int a[N][N]; int st[N][N][2]; //st[i][j][k]表示从(i,j)开始,大小为2^k的三角形内最大值,第三维用滚动数组优化 int n, k; long long ans; int q[N], head, tail; int main() { cin >> n >> k; for(int i = 1; i <= n; i++) { for(int j = 1; j <= i; j++) { cin >> a[i][j]; st[i][j][0] = a[i][j]; } } for(int l = 1; l <= log2(k); l++) { for(int i = 1; i <= n; i++) { head = 1, tail = 0; //单调队列头、尾指针初始化 int len = (1 << l - 1) + 1; //窗口大小 int x = i + (1 << l - 1); //起始横坐标 for(int j = 1; j <= n; j++) { st[i][j][l % 2] = st[i][j][(l - 1) % 2]; //去头 while(head <= tail && j - q[head] + 1 > len) head++; //去尾 while(head <= tail && st[x][j][(l - 1) % 2] > st[x][q[tail]][(l - 1) % 2]) tail--; //存入 q[++tail] = j; //取窗口最大值,并与i,j - len + 1位置对应的值取max if(j >= len) st[i][j - len + 1][l % 2] = max(st[i][j - len + 1][l % 2], st[x][q[head]][(l - 1) % 2]); } } } for(int i = 1; i + k - 1 <= n; i++) { for(int j = 1; j <= i; j++) { int len = log2(k); int maxijk = st[i][j][len % 2]; int x = i + k - (1 << len); for(int y = j; y <= j + k - (1 << len); y++) { maxijk = max(maxijk, st[x][y][len % 2]); } ans += maxijk; } } cout << ans; return 0; }
10-02
根据原作 https://pan.quark.cn/s/459657bcfd45 的源码改编 Classic-ML-Methods-Algo 引言 建立这个项目,是为了梳理和总结传统机器学习(Machine Learning)方法(methods)或者算法(algo),和各位同仁相互学习交流. 现在的深度学习本质上来自于传统的神经网络模型,很大程度上是传统机器学习的延续,同时也在不少时候需要结合传统方法来实现. 任何机器学习方法基本的流程结构都是通用的;使用的评价方法也基本通用;使用的一些数学知识也是通用的. 本文在梳理传统机器学习方法算法的同时也会顺便补充这些流程,数学上的知识以供参考. 机器学习 机器学习是人工智能(Artificial Intelligence)的一个分支,也是实现人工智能最重要的手段.区别于传统的基于规则(rule-based)的算法,机器学习可以从数据中获取知识,从而实现规定的任务[Ian Goodfellow and Yoshua Bengio and Aaron Courville的Deep Learning].这些知识可以分为四种: 总结(summarization) 预测(prediction) 估计(estimation) 假想验证(hypothesis testing) 机器学习主要关心的是预测[Varian在Big Data : New Tricks for Econometrics],预测的可以是连续性的输出变量,分类,聚类或者物品之间的有趣关联. 机器学习分类 根据数据配置(setting,是否有标签,可以是连续的也可以是离散的)和任务目标,我们可以将机器学习方法分为四种: 无监督(unsupervised) 训练数据没有给定...
本系统采用微信小程序作为前端交互界面,结合Spring Boot与Vue.js框架实现后端服务及管理后台的构建,形成一套完整的电子商务解决方案。该系统架构支持单一商户独立运营,亦兼容多商户入驻的平台模式,具备高度的灵活性与扩展性。 在技术实现上,后端以Java语言为核心,依托Spring Boot框架提供稳定的业务逻辑处理与数据接口服务;管理后台采用Vue.js进行开发,实现了直观高效的操作界面;前端微信小程序则为用户提供了便捷的移动端购物体验。整套系统各模块间紧密协作,功能链路完整闭环,已通过严格测试与优化,符合商业应用的标准要求。 系统设计注重业务场景的全面覆盖,不仅包含商品展示、交易流程、订单处理等核心电商功能,还集成了会员管理、营销工具、数据统计等辅助模块,能够满足不同规模商户的日常运营需求。其多店铺支持机制允许平台方对入驻商户进行统一管理,同时保障各店铺在品牌展示、商品销售及客户服务方面的独立运作空间。 该解决方案强调代码结构的规范性与可维护性,遵循企业级开发标准,确保了系统的长期稳定运行与后续功能迭代的可行性。整体而言,这是一套技术选型成熟、架构清晰、功能完备且可直接投入商用的电商平台系统。 资源来源于网络分享,仅用于学习交流使用,请勿用于商业,如有侵权请联系我删除!
评论
添加红包

请填写红包祝福语或标题

红包个数最小为10个

红包金额最低5元

当前余额3.43前往充值 >
需支付:10.00
成就一亿技术人!
领取后你会自动成为博主和红包主的粉丝 规则
hope_wisdom
发出的红包
实付
使用余额支付
点击重新获取
扫码支付
钱包余额 0

抵扣说明:

1.余额是钱包充值的虚拟货币,按照1:1的比例进行支付金额的抵扣。
2.余额无法直接购买下载,可以购买VIP、付费专栏及课程。

余额充值