微行为视角：电商推荐系统的新进展

需积分: 9 14 浏览量更新于2024-09-07 收藏 1.76MB PDF 举报

"2018年的研究论文《MicroBehaviors: A New Perspective in E-commerce Recommender Systems》提出了电子商务推荐系统的新视角。随着电子商务网站的爆炸性增长，用户的购物习惯发生了深刻变化，越来越多的人倾向于在线购物，这为商家提供了丰富的用户数据。传统推荐系统主要关注用户与商品的宏观交互，即用户的购买历史，但这篇论文强调了用户与商品之间的微观行为（MicroBehaviors）的重要性。微观行为是指在用户与商品的一次宏观交互过程中，用户对商品进行的具体操作序列，如浏览、阅读评论、加入购物车以及实际下单等。这些行为反映了用户对商品的详细关注度和兴趣深度，提供了对用户行为的精细洞察。通过分析这些微小但关键的行为，研究人员能够更深入地理解用户需求和偏好，从而为推荐系统的发展创造巨大机会。论文作者Meizi Zhou、Zhuoye Ding、Jiliang Tang和Dawei Yin来自不同的学术机构，他们指出，除了传统的宏观交互数据，挖掘微观行为数据可以帮助推荐算法更好地个性化推荐，提升用户体验，提高转化率，并可能开启新的推荐策略。例如，通过识别用户的浏览路径和停留时间，推荐系统可以预测用户可能的兴趣走向，甚至在用户还未明确表达需求时就提前推荐相关商品。这篇文章提倡将微观行为作为电子商务推荐系统的一个重要补充，以实现更加精准和动态的个性化推荐，这是在大数据时代背景下对推荐系统理论和技术的重要拓展。这种新视角对于优化用户购物体验，提升电商行业的竞争力具有深远影响。"

Micro Behaviors: A New Perspe ctive in E-commerce

Recommender Systems

Meizi Zhou

∗

University of Minnesota

zhou0793@umn.edu

Zhuoye Ding

Data Science Lab, JD.com

dingzhuoye@jd.com

Jiliang Tang

Michigan State University

tangjili@msu.edu

Dawei Yin

†

Data Science Lab, JD.com

yindawei@acm.org

ABSTRACT

The explosive popularity of e-commerce sites has reshaped users’

shopping habits and an increasing number of users prefer to spend

more time shopping online. This evolution allows e-commerce sites

to observe rich data about users. The majority of traditional recom-

mender systems have focused on the macro interactions between

users and items, i.e., the purchase history of a customer. However,

within each macro interaction between a user and an item, the user

actually performs a sequence of micro behaviors, which indicate

how the user locates the item, what activities the user conducts

on the item (e.g., reading the comments, carting, and ordering)

and how long the user stays with the item. Such micro behaviors

oer ne-grained and deep understandings about users and pro-

vide tremendous opportunities to advance recommender systems

in e-commerce. However, exploiting micro behaviors for recom-

mendations is rather limited, which motivates us to investigate e-

commerce recommendations from a micro-behavior perspective in

this paper. Particularly, we uncover the eects of micro behaviors on

recommendations and propose an interpretable

ecommendation

framework RIB, which models inherently the sequence of m

cro

ehaviors and their eects. Experimental results on datasets from a

real e-commence site demonstrate the eectiveness of the proposed

framework and the importance of micro behaviors for recommen-

dations.

CCS CONCEPTS

• Information systems → Summarization;

KEYWORDS

Micro behaviors, RNN, attention mechanism, e-commerce, recom-

mendation

∗

This work was done, when the author was an internship at Data Science Lab of

JD.com.

†

Corresponding author

Permission to make digital or hard copies of all or part of this work for personal or

classroom use is granted without fee provided that copies are not made or distributed

for prot or commercial advantage and that copies bear this notice and the full citation

on the rst page. Copyrights for components of this work owned by others than ACM

must be honored. Abstracting with credit is permitted. To copy otherwise, or republish,

to post on servers or to redistribute to lists, requires prior specic permission and/or a

fee. Request permissions from permissions@acm.org.

WSDM 2018, February 5–9, 2018, Marina Del Rey, CA, USA

ACM ISBN 978-1-4503-5581-0/18/02.. . $15.00

https://doi.org/10.1145/3159652.3159671

ACM Reference Format:

Meizi Zhou, Zhuoye Ding, Jiliang Tang, and Dawei Yin. 2018. Micro Be-

haviors: A New Perspective in E-commerce Recommender Systems. In

Proceedings of WSDM 2018: The Eleventh ACM International Conference on

Web Search and Data Mining (WSDM 2018). ACM, New York, NY, USA,

9 pages. https://doi.org/10.1145/3159652.3159671

1 INTRODUCTION

The modern e-commerce sites such as Amazon

and eBay

oer

hundreds of millions of products for sale. For example, as on June

20th, 2017, Amazon has more than 372 million products

. It has

become increasingly challenging for consumers to nd their inter-

ested items. Recommender systems play a crucial role in mitigating

this information overload problem by suggesting products that

have potentials to t consumers’ needs. They have been proven to

not only help increase customer satisfaction and create customer

loyalty [

] but also boost many aspects of e-commerce services

such as revenue and growth [24].

Meanwhile, the popularity of e-commerce sites has reshaped

users’ shopping habits and users prefer to spend more time shop-

ping online. For example, on average, American parents spend 7

hours per week on e-commerce sites

. This evolution enables e-

commerce sites to observe rich data about their users. Figure 1

illustrates a real example of observed data on a user from an e-

commerce site in a short period. The user rst enters a page of

iPhone 7 from searching result page. She reads the detailed descrip-

tion, as well as others’ comments and adds it to the cart. Then she

shifts to a page of iPhone 6 from the searching result page and reads

the comments. After that, she browses a page of iPhone 7 cases

from the sale page and orders the case. Finally, she jumps to a page

of Samsung Galaxy from the home page of the e-commerce site.

From a macro perspective as shown in the top subgure in Figure 1,

the user interacted with iPhone 7, iPhone 6, iPhone 7 cases and

Samsung Galaxy. While from a micro perspective as shown in the

bottom subgure, each macro interaction includes a sequence of

behaviors that can indicate how the user located the product page

(e.g., the search engine or the sale promotion), whether the user

clicks detailed information about a product (e.g., comments, or spec-

ications), whether a user carts or orders a product, and how long

the user dwells on a product. In this work, we refer these behaviors

https://www.amazon.com/

https://www.ebay.com/

https://www.scrapehero.com/number-of-products-sold-on-amazon-com-june-2017/

https://www.bigcommerce.com/blog/ecommerce-trends/#cmtoc_anchor_id_0

Technical Presentation

WSDM’18, February 5-9, 2018, Marina Del Rey, CA, USA

727

下载后可阅读完整内容，剩余8页未读，立即下载

Jayxp

粉丝: 6
资源: 137

微行为视角：电商推荐系统的新进展

100万条京东评分数据集

Lerner -- Python Workout. 50 Essential Exercises -- 2020.pdf

Phase control of optical steady-state behaviors from Fano-type interference in triple-semiconductor quantum wells

Brand Purchase Prediction based on Time-evolving User Behaviors in E-commerce

论文研究-Brand Purchase Prediction based on Time-evolving User Behaviors.pdf

Teaching peer-tutoring behaviors in first- and third-grade classrooms

Python库 | collective.behaviors-1.1.1.tar.gz

Long-term effects of a token economy on target and off-task behaviors

Tom.Mitchell.-.Machine.Learning

Detecting Application-layer Attacks Based on User's Application-layer Behaviors

最新资源