Post-Click Conversion Rate Predictive Model on E-commerce Recommender System

Public Deposited
Analytics

Downloadable Content

Download PDF
Request Version for Screen Reader
Last Modified
  • February 28, 2019
Creator
  • Ding, Yuhe
    • Affiliation: School of Information and Library Science
Abstract
  • This paper discusses about how different features influence customers’ decision on their online purchase after click behavior. The dataset is gathered from real-world traffic log of the recommender system in e-commerce. Logistic Regression and Extreme Gradient Boosting are used as main machine learning approaches for predictive analysis and modeling. In this study, features from users’ profile, shops’ profile and context are tested to see to what extent they may exert influence on customers’ purchase intention. Based on the experiment results, this paper also proposes some possible improvement for e-commerce platform in personalized recommendation in order to increase conversions and discusses about potential approaches to improve conversion rate prediction performance.
Date of publication
Keyword
DOI
Resource type
Advisor
  • Shaw, Ryan
Degree
  • Master of Science in Information Science
Academic concentration
  • Library and Information Science
Degree granting institution
  • University of North Carolina at Chapel Hill
Graduation year
  • 2018
Language
Deposit record
  • 9023c49e-238b-40d7-9235-640284603c3e
Date uploaded
  • December 12, 2018

Relations

Parents:

This work has no parents.

In Collection:

Items

Thumbnail Title Date Uploaded Visibility Actions
file details yuheding.pdf December 12, 2018 Public Download
file details Parameter Tunning.ipynb December 12, 2018 Public Download
file details Feature Engineering.ipynb December 12, 2018 Public Download
file details XGboost.ipynb December 12, 2018 Public Download
file details Data Preprocessing.ipynb December 12, 2018 Public Download
file details XGBoost+lr.ipynb December 12, 2018 Public Download