o2o

本次赛题为《生活大实惠：O2O优惠券使用预测》，是新人赛提供的第二个赛题

本赛题的比赛背景：随着移动设备的完善和普及，移动互联网+各行各业进入了高速发展阶段，这其中以O2O（Online to Offline）消费最为吸引眼球。据不完全统计，O2O行业估值上亿的创业公司至少有10家，也不乏百亿巨头的身影。O2O行业天然关联数亿消费者，各类APP每天记录了超过百亿条用户行为和位置记录，因而成为大数据科研和商业化运营的最佳结合点之一。 以优惠券盘活老用户或吸引新客户进店消费是O2O的一种重要营销方式。然而随机投放的优惠券对多数用户造成无意义的干扰。对商家而言，滥发的优惠券可能降低品牌声誉，同时难以估算营销成本。 个性化投放是提高优惠券核销率的重要技术，它可以让具有一定偏好的消费者得到真正的实惠，同时赋予商家更强的营销能力。本次大赛为参赛选手提供了O2O场景相关的丰富数据，希望参赛选手通过分析建模，精准预测用户是否会在规定时间内使用相应优惠券。

赛制说明

本场比赛长期开放，报名和参赛无时间限制。排行榜三个月更新一次

参赛报名

1. 要求以个人形式参与比赛，并确保报名信息准确有效；

2. 报名方式：用淘宝或阿里云账号登入天池官网，完成个人信息注册，即可报名参赛；

3.学习文档&培训课程的链接

参赛规则

1. 报名成功后，选手下载数据，在本地调试算法，提交结果；

2. 提交后将进行实时评测；每天排行榜更新时间为12:00和20:00，按照评测指标得分从高到低排序；排行榜将选择历史最优成绩进行展示；

参赛对象

大赛面向全社会开放，参赛对象不限，要求以个人形式参赛。

请注意提交的数据格式(Table 4和sample_submission.csv)。

数据
本赛题提供用户在2016年1月1日至2016年6月30日之间真实线上线下消费行为，预测用户在2016年7月领取优惠券后15天以内的使用情况。
注意： 为了保护用户和商家的隐私，所有数据均作匿名处理，同时采用了有偏采样和必要过滤。

评价方式
本赛题目标是预测投放的优惠券是否核销。针对此任务及一些相关背景知识，使用优惠券核销预测的平均AUC（ROC曲线下面积）作为评价标准。 即对每个优惠券coupon_id单独计算核销预测的AUC值，再对所有优惠券的AUC值求平均作为最终的评价标准。 关于AUC的含义与具体计算方法，可参考维基百科

This competition is titled *Great Deals in Life: O2O Coupon Usage Prediction*, and it is the second problem provided for the newcomer contest.

### Competition Background
With the improvement and popularization of mobile devices, mobile internet integrated with various industries has entered a stage of rapid development, among which O2O (Online to Offline) consumption stands out the most. According to incomplete statistics, there are at least 10 startups in the O2O industry with a valuation of over 100 million RMB, and there are also giants with valuations exceeding 10 billion RMB. The O2O industry naturally connects hundreds of millions of consumers, and various apps record over 10 billion user behavior and location records daily, making it one of the best combinations of big data research and commercial operations.

Using coupons to revitalize existing users or attract new customers to shop offline is an important marketing method for O2O. However, randomly distributed coupons cause meaningless interference to most users. For merchants, over-issued coupons may damage brand reputation and make it difficult to estimate marketing costs. Personalized coupon delivery is an important technology to improve the coupon redemption rate, which can provide real benefits to consumers with specific preferences, while endowing merchants with stronger marketing capabilities. This competition provides rich data related to O2O scenarios for participants, expecting them to conduct analysis and modeling to accurately predict whether users will use the corresponding coupons within the specified time frame.

### Competition Rules
This competition is open for a long term, with no time limits for registration and participation. The leaderboard is updated every three months.

### Registration Requirements
1. Participants must compete individually and ensure that their registration information is accurate and valid;
2. Registration method: Log in to the official Tianchi website using a Taobao or Alibaba Cloud account, complete personal information registration, and then you can register for the competition;
3. Links to learning documents & training courses

### Competition Participation Rules
1. After successful registration, participants may download the data, debug algorithms locally, and submit results;
2. Real-time evaluation will be conducted after submission; the leaderboard is updated at 12:00 and 20:00 every day, and sorted by the evaluation indicator scores in descending order; the leaderboard will display the historical best performance of each participant;
Participants are requested to pay attention to the required submission data format (Table 4 and sample_submission.csv).

### Eligibility for Participation
This competition is open to the whole society, with no restrictions on participants, and individual participation is required.

### Data
This competition provides users' real online and offline consumption behaviors between January 1, 2016 and June 30, 2016, and requires participants to predict the usage status of coupons received by users within 15 days after they claim the coupons in July 2016.
Note: To protect the privacy of users and merchants, all data has been anonymized, and biased sampling and necessary filtering have been applied.

### Evaluation Metric
The goal of this competition is to predict whether a distributed coupon will be redeemed. For this task and related background knowledge, the average AUC (Area Under the ROC Curve) of coupon redemption prediction is used as the evaluation standard. Specifically, the AUC value of redemption prediction is calculated individually for each coupon_id, then the average of all coupon AUC values is taken as the final evaluation standard. For the meaning and specific calculation method of AUC, please refer to Wikipedia.

搜集汇总

背景与挑战

背景概述

该数据集为O2O优惠券使用预测赛题提供的数据，包含2016年1月1日至6月30日的用户线上线下消费行为，用于预测用户在7月领取优惠券后15天内的使用情况。数据经过匿名和有偏采样处理，评价标准为优惠券核销预测的平均AUC。