首页 > AI前沿 > The Window Lever: Longer Observation Windows Outweigh the Observed Effects of Data, Capacity, and Architecture in Customer Behavior Forecasting

The Window Lever: Longer Observation Windows Outweigh the Observed Effects of Data, Capacity, and Architecture in Customer Behavior Forecasting

arXiv机器学习 2026-11-20 02:00 9 阅读 查看原文
用户现在需要把给定的新闻正文排版成语义化的HTML片段,首先得先分段落对吧?首先看原文的段落,首先第一个自然段是讲实验里长观测窗口的作用对吧?不对,先拆分每个自然段落: 首先第一部分:从开头“In our experiments on long-horizon customer-behavior forecasting”到“longer windows and do not identify the magnitude of a calendar-independent history effect.”?不对,仔细读原文的分段,哦原文是连续的,我要把每个自然段拆对?不对,先理清楚逻辑: 哦不对,原文的内容: 第一个段落:In our experiments on long-horizon customer-behavior forecasting, a longer observation window outweighed the observed effects of more data, more capacity and newer architectures. The window is the weeks of history a model sees before the week it forecasts from. Going from an 8- to a 16-week window adds a raw +0.070 to 0.107 log-space $R^2$ in every family at 1M. Matched test cohorts, which score the same customer-weeks at every window, show this raw comparison understates the gain in every family they cover. The direction holds on five public datasets, four spanning over a year (ridge and our transformer on all five). On our internal data, longer windows admit up to 10% more and colder training customers, which matching does not fix. Our design does not separate window length from calendar coverage, because every test anchor (the week a forecast is made from) falls in one eight-week band. On training-period anchors, the weekly ridge probe's gain persists outside the holiday season, keeping about 72% of it. We therefore report a robust positive association with longer windows and do not identify the magnitude of a calendar-independent history effect. 然后第二个段落:Of four internally registered predictions, the only fully blind one failed. An activity-mixture model predicted +0.014 from 24 to 32 weeks; the weekly-aggregate probe returned -0.0004. 第三个段落:For the sequence families, the other three levers are about an order of magnitude smaller, bounded and unresolved. Their observed effects are at most 0.006 between the nominal 1M and 10M tiers and 0.0052 among five sequence families at 16 weeks with clocks as built, both single-seed on the full test split, and 0.006 for capacity on three-seed means. These are observed intervals, not equivalence bounds. Our claim is the order-of-magnitude contrast between the window and those bounds. 第四个段落:A team building on this class of data should start by lengthening the window. 不对?或者有没有需要用h2/h3的?哦规则说长文用h2/h3划分小节,那这里可以划分几个小节?比如第一个小节可以叫

观测窗口的核心实验结论

?不对不对,不能增删内容啊!哦对,规则4是保留原文所有信息,不得增删改内容!哦对,不能加自己写的标题内容对吧?哦不对,等下,h2/h3是划分小节,但是标题内容不能自己加?不对不对,原文里有没有小标题?没有,那原文是几个自然段落?等下再仔细看用户的要求:3.每个自然段落用p包裹,长文用h2/h3划分小节,但是不能增删改内容,哦,不对,是不是如果原文本身有层次的话可以划分,但是不能自己加额外的文字?哦不对,等下原文的内容,有没有可以划分的?哦不对,我刚才想错了,不能加原文没有的文字,那如果原文没有小标题,那是不是不用h2/h3?不对不对,再看规则:长文用h2/h3划分小节,但是不能增删内容,哦,哦可能是把内容按照逻辑划分,但是h2/h3的内容也要是原文里的?不对,原文里没有小标题,那是不是就都用p?不对,等下再仔细读原文,哦,不对,有没有引用、列表、代码?原文里没有列表,没有引用,没有代码,那重要的句子可以用strong。比如“longer observation window outweighed the observed effects of more data, more capacity and newer architectures”可以加strong?还有“robust positive association with longer windows”可以加strong?还有“order-of-magnitude contrast between the window and those bounds”可以加strong?还有最后一句“A team building on this class of data should start by lengthening the window.”可以加strong? 等下现在整理: 首先第一个自然段,整个是一段,用p包裹,里面的重点句子加strong:

In our experiments on long-horizon customer-behavior forecasting, a longer observation window outweighed the observed effects of more data, more capacity and newer architectures. The window is the weeks of history a model sees before the week it forecasts from. Going from an 8- to a 16-week window adds a raw +0.070 to 0.107 log-space $R^2$ in every family at 1M. Matched test cohorts, which score the same customer-weeks at every window, show this raw comparison understates the gain in every family they cover. The direction holds on five public datasets, four spanning over a year (ridge and our transformer on all five). On our internal data, longer windows admit up to 10% more and colder training customers, which matching does not fix. Our design does not separate window length from calendar coverage, because every test anchor (the week a forecast is made from) falls in one eight-week band. On training-period anchors, the weekly ridge probe's gain persists outside the holiday season, keeping about 72% of it. We therefore report a robust positive association with longer windows and do not identify the magnitude of a calendar-independent history effect.

然后第二个自然段:Of four internally registered predictions, the only fully blind one failed. An activity-mixture model predicted +0.014 from 24 to 32 weeks; the weekly-aggregate probe returned -0.0004. 这里有没有什么重点?哦“the only fully blind one failed”可以加?或者不用,也可以,不过规则说重要句子可用,不是必须。

Of four internally registered predictions, the only fully blind one failed. An activity-mixture model predicted +0.014 from 24 to 32 weeks; the weekly-aggregate probe returned -0.0004.

第三个自然段:For the sequence families, the other three levers are about an order of magnitude smaller, bounded and unresolved. Their observed effects are at most 0.006 between the nominal 1M and 10M tiers and 0.0052 among five sequence families at 16 weeks with clocks as built, both single-seed on the full test split, and 0.006 for capacity on three-seed means. These are observed intervals, not equivalence bounds. Our claim is the order-of-magnitude contrast between the window and those bounds. 哦对,这里的claim是重点,加strong。然后这个段落:

For the sequence families, the other three levers are about an order of magnitude smaller, bounded and unresolved. Their observed effects are at most 0.006 between the nominal 1M and 10M tiers and 0.0052 among five sequence families at 16 weeks with clocks as built, both single-seed on the full test split, and 0.006 for capacity on three-seed means. These are observed intervals, not equivalence bounds. Our claim is the order-of-magnitude contrast between the window and those bounds.

然后最后一个自然段:

A team building on this class of data should start by lengthening the window.

不对?等下有没有需要h2/h3的?哦原文有没有长文的层次?哦原文的内容是先讲窗口的核心结论,然后讲预测验证,然后讲其他杠杆的效果,然后给建议,但是不能自己加标题内容啊,因为规则说不得增删改内容,加h2的标题文字就相当于加了原文没有的内容了对吧?对,所以不能自己编小标题,所以就不用h2/h3了,因为原文本身没有小节标题,所有都是自然段落,所以都用p包裹,重点句子加strong就可以了? 等下再检查有没有漏内容:所有原文的内容都保留了对吧?对,每个自然段都分开了,重点句子加strong,没有加任何额外内容,没有文档级标签,符合要求。 哦等下再仔细看原文的每个句子,有没有拆分错误?比如第一个自然段是不是确实是一个自然段?对,原文从开头到“calendar-independent history effect”是一个连续