Overview
Topic status: In progress
Overview
As information overload problem has been a serious issue for Web users, how to help users deal with this problem becomes very important. Recently, personalised recommender systems have been developed to overcome the information overload problem and provide personalised content, services and information items to potential consumers. The goal of this project is to develop a Java program to implement a tag recommender system by using Mahout framework. Mahout is a popular Java based machine learning library, it provides a rich set of machine learning algorithms as well as recommender related algorithms.
For more information about Mahout visit the Apache Mahout website.
By utilising the library, it is expected that students will:
- avoid reinventing the wheel, and hence reduce the time and efforts on implementation, prototyping and evaluation of the system
- improve coding and implementation quality by following concepts, styles and architectures of the state-of-art machine-learning library (i.e. the Mahout)
- potentially extend the algorithm implementation or evaluation process into distributed fashion by utilising the Hadoop framework
Research activities
The following activities are involved in this project:
- read articles about recommender systems and related areas in order to obtain the knowledge of recommendation making techniques
- learn how to use the Mahout framework and MongoDB database
- develop Java programs to implement tag recommendation algorithms in the Mahout environment
- test and evaluate the developed system using a given dataset.
Expected outcomes
The expected outcomes for this topic are:
- a database in MongoDB to store the data in a given dataset
- Java programs implementing given tag recommendation algorithms
- a report of testing result analysis and evaluation.
Duration
8 weeks from 2 January to 24 February 2012.
- Study level
- Vacation research experience scholarship
- Supervisors
- QUT
- Organisational unit
Science and Engineering Faculty
- Research area
- Contact
- Please contact supervisor for enquiries.
Associate Professor Yue Xu
