Context-aware Implicit Feedback Recommendation based on Repeat Consumption Behaviors

  • 楊 璨瑜

Student thesis: Master's Thesis


The importance of recommendation systems is increasing since the rise of the e-commerce It is an efficient way to retrieve wanted data for customers within flood of information An accurate and efficient recommendation system is helps users to find the product they want more conveniently which improving the revenue of the e-commerce platform by providing information of interest to customers However how to retrieve and use the customer’s using record for building a reliable recommendation system has become an important issue There are two kinds of user feedback: explicit feedback and implicit feedback Traditional recommendation methods collected numeric scores on products as customers’ feedback The score data is called as explicit feedback However this kind of rating score data is rare and very sparse Therefore other researches focus on using the original customer records as recommended basis is called implicit feedback Nonetheless implicit feedback cannot reflect the customer preferences directly that cause worse performance in some cases In this study we proposed a context-aware recommendation approach to solve the problem of recommendation on implicit feedback This approach considers the features that identify context extracted from customers’ behaviors especially on the repeat consumption behaviors As evaluation we compare our approach with other implicit feedback recommendation approaches context-aware recommendation approaches are also compared Our approach can be applied to several domains e g webpage viewing music listening query searching and usage habits on intelligent devices
Date of Award2016 Aug 30
Original languageEnglish
SupervisorHung-Yu Kao (Supervisor)

Cite this