Ecosyste.ms: Awesome

An open API service indexing awesome lists of open source software.

Awesome Lists | Featured Topics | Projects

https://github.com/mjackie/recsys

计算广告/推荐系统/机器学习(Machine Learning)/点击率(CTR)/转化率(CVR)预估/点击率预估
https://github.com/mjackie/recsys

ctr ctr-prediction machine-learning recommendation-system recommender-system

Last synced: 30 days ago
JSON representation

计算广告/推荐系统/机器学习(Machine Learning)/点击率(CTR)/转化率(CVR)预估/点击率预估

Host: GitHub
URL: https://github.com/mjackie/recsys
Owner: mJackie
Created: 2019-03-08T14:16:10.000Z (over 5 years ago)
Default Branch: master
Last Pushed: 2019-12-17T10:44:07.000Z (almost 5 years ago)
Last Synced: 2024-10-14T11:21:05.396Z (30 days ago)
Topics: ctr, ctr-prediction, machine-learning, recommendation-system, recommender-system
Homepage:
Size: 42.1 MB
Stars: 1,943
Watchers: 70
Forks: 437
Open Issues: 1
Metadata Files:
- Readme: README.md

Awesome Lists containing this project

README

        推荐系统/计算广告/机器学习/CTR预估资料汇总

===

[![License](https://img.shields.io/badge/license-MIT-blue.svg)](./LICENSE.md) ![Update](https://img.shields.io/badge/update-weekly-green.svg) ![Progress](https://img.shields.io/badge/progress-1003%20%2F%201003-ff69b4.svg) [![SayThanks](https://img.shields.io/badge/say-thanks-ff69f4.svg)](https://saythanks.io/to/kamyu104) ![Travis](https://travis-ci.org/kamyu104/LeetCode-Solutions.svg?branch=master)

**说明**：本仓库主要汇集推荐系统/计算广告/机器学习/CTR预估相关学习资料，欢迎一起补充更新👼

**更多前沿技术文章，欢迎移步 -> [精选文章](https://github.com/mJackie/RecNews)**

# RoadMap

- [推荐系统](#推荐系统)

- [计算广告](#计算广告)

- [统计学习模型](#统计学习模型)

	- [技术文章](#技术文章)

	- [实践工具](#实践工具)

- [深度学习模型](#深度学习模型)

	- [技术文章](#技术文章)

	- [实践代码](#实践代码)

- [相关比赛](#相关比赛)

	- [Criteo Display Advertising Challenge](#CriteoDisplayAdvertisingChallenge)

	- [Avazu Click Through Rate Prediction](#AvazuClickThroughRatePrediction)

	- [2018 IJCAI 阿里妈妈搜索广告转化预测](#2018-IJCAI-阿里妈妈搜索广告转化预测)

	- [2018腾讯广告算法大赛](#2018腾讯广告算法大赛)

- [技术博客](#技术博客)

- [经典论文清单](#经典论文清单)

---

### 推荐系统

- [推荐系统简明教程-概述](https://zhuanlan.zhihu.com/p/87411668)

- [推荐系统简明教程-召回](https://zhuanlan.zhihu.com/p/87578318)

- [推荐系统简明教程-排序](https://zhuanlan.zhihu.com/p/87796986)

- [推荐系统实践-项亮](./resource/推荐系统实践-项亮.pdf)

---

### 计算广告

- [互联网广告系统综述系列博文](https://blog.csdn.net/mytestmy/article/list)

- [计算广告学系列视屏-刘鹏](https://study.163.com/course/introduction.htm?courseId=321007#/courseDetail?tab=1)

- [计算广告学讲义-刘鹏](https://dirtysalt.github.io/html/computational-advertising.html)

---

### 统计学习模型

#### 技术文章

- [前深度学习时代CTR预估模型的演化之路](https://zhuanlan.zhihu.com/p/61154299)

- [逻辑回归LR模型简介](https://tech.meituan.com/2015/05/08/intro-to-logistic-regression.html)

- [FFM讲解PPT](./resource/ffm.pdf)

- [深入FFM原理与实践](https://tech.meituan.com/2016/03/03/deep-understanding-of-ffm-principles-and-practices.html)

- [GBDT算法原理与系统设计简介](./resource/GBDT-wepon.pdf)

#### 实践工具

- [LightGBM](https://github.com/Microsoft/LightGBM)

- [XGBoost](https://github.com/dmlc/xgboost)

- [LIBFFM](https://github.com/guestwalk/libffm)

- [xLearn](https://github.com/aksnzhy/xlearn)

- [DeepCTR](https://github.com/shenweichen/DeepCTR)

---

### 深度学习模型

#### 技术文章

- [深度学习如何应用在广告、推荐及搜索业务？](https://mp.weixin.qq.com/s/nboZ6p_l30L__FJNyz6Ohw)

- [深度学习在CTR预估中的应用](https://zhuanlan.zhihu.com/p/35484389)

- [深度学习在 CTR 中应用](http://www.mamicode.com/info-detail-1990002.html)

- [深度学习在美团点评推荐业务中实践](https://gitbook.cn/books/5aa0dd15cfbe2c144b71906d/index.html)

#### 实践代码

- [CTR预估算法之FM, FFM, DeepFM及实践](https://github.com/Johnson0722/CTR_Prediction)

- [推荐系统中使用ctr排序的dnn模型](https://github.com/nzc/dnn_ctr)

---

### 相关比赛

#### [CriteoDisplayAdvertisingChallenge](https://www.kaggle.com/c/criteo-display-ad-challenge)

- Rank1: [借鉴了Facebook的方案: GBDT 特征编码 + FFM](https://www.kaggle.com/c/criteo-display-ad-challenge/discussion/10555)

- Rank3: [Quadratic Feature Generation + FTRL 传统特征工程和 FTRL 线性模型的结合](https://www.kaggle.com/c/criteo-display-ad-challenge/discussion/10534)

#### [AvazuClickThroughRatePrediction](https://www.kaggle.com/c/avazu-ctr-prediction)

- Rank1: [Feature Engineering + FFM + Ensemble, 只基于 FFM 进行集成](https://www.kaggle.com/c/avazu-ctr-prediction/discussion/12608)

- Rank2: [Feature Engineering + GBDT 特征编码 + FFM + Blending](https://github.com/owenzhang/kaggle-avazu)

#### [2018 IJCAI 阿里妈妈搜索广告转化预测](https://tianchi.aliyun.com/competition/entrance/231647/introduction?spm=5176.12281957.1004.10.38b04c2aaROEf9)

- Rank1: [GBDT，用了嫁接技术处理样本分布不一致](https://github.com/plantsgo/ijcai-2018)

- Rank2: [GBDT单模型+大量特征工程](https://github.com/YouChouNoBB/ijcai-18-top2-single-mole-solution)

#### [2018腾讯广告算法大赛](https://algo.qq.com/)

- Rank3: https://github.com/DiligentPanda/Tencent_Ads_Algo_2018

- Rank6: https://github.com/nzc/tencent-contest

---

### 技术博客

- [深度推荐系统](https://zhuanlan.zhihu.com/deep-recsys)

- [美团技术点评](https://tech.meituan.com/)

- [深度学习沿言笔记](https://zhuanlan.zhihu.com/c_188941548)

- [王喆的机器学习笔记](https://zhuanlan.zhihu.com/wangzhenotes)

- [推荐系统理论及实战](https://www.jianshu.com/nb/21403842)

- [刘十三的机器学习笔记](https://zhuanlan.zhihu.com/c_1169669630389440512)

---

### 经典论文清单

筛选文章的标准：**前沿或者经典的，工程导向的，google、阿里、facebook等一线互联网公司出品的**：

[**Wide & Deep Learning for Recommender Systems**](./paper/wide&deep.pdf)

> google 的 wide&deep，必看论文，经典到难以附加

[**DeepFM: An End-to-End Wide & Deep Learning

Framework for CTR Prediction**](./paper/deepfm.pdf)

> 华为对wide&deep的改进，加了wide层的交叉项。如今工业界的主流模型

[**Practical lessons from predicting clicks on ads at facebook**](./paper/2014GBDT+LR-Facebook.pdf)

> facebook GBDT+LR的经典方案。虽然如今已不是主流方案，但论文中的思想很值得学习。

[**Deep Neural Networks for YouTube Recommendations**](./paper/youtube-recsys.pdf)

> 介绍了Youtube推荐系统工业界架构与方案，经典必看

[**Real-time Personalization using Embeddings for Search Ranking at Airbnb**](./paper/airbnb2018KDD.pdf)

> KDD2018 best paper，Embedding 必看论文，非常经典

[**Entire Space Multi-Task Model: An Effective Approach for Estimating Post-Click Conversion Rate**](./paper/ESSM2018.pdf)

> 阿里的多目标学习经典方案，同时优化CTR & CVR

[**Real-time Personalization using Embeddings for Search Ranking at Airbnb**](./paper/aribnbSearch.pdf)

> 介绍了 airbnb 搜索排序模型的演进，工业性质很强，值得参考

[**搜索引擎点击模型综述**](./paper/搜索引擎点击模型综述.pdf)

> 清华马少平团队的文章点击模型入门必看，搜索引擎点击模型综述