Personalized Preference-based Service Selection and Recommendation

Kenneth K. Fletcher

The following Great Innovative Idea is from Kenneth K. Fletcher, an Assistant Professor of Computer Science at the University of Massachusetts Boston. Fletcher presented his poster, A Method for Dealing with Data Sparsity and Cold-Start Limitations in Service Recommendation Using Personalized Preferences, at the CCC Symposium on Computing Research, October 23-24, 2017.

The Idea

The idea is to develop and improve on methods and algorithms that incorporate users’ preferences for service selection and recommendation. With the growing number of services (cloud, web, airline, etc) on the Internet having similar functionalities, it is now more challenging to select or recommend the best service(s) that meet(s) users’ needs. Rather than selecting or recommending services based only on its functionality, users are increasingly paying more attention to non-functional attributes (like availability, throughput, cost, etc) because they provide a distinction among the competing services with similar functionality. Using non-functional attributes, however, presents service providers with personalization and conflicting non-functional attributes challenges. These challenges create a gap between users’ non-functional attribute values and their satisfaction with service selection or recommendation. To bridge this gap and improve service selection and recommendation accuracy, it is important to incorporate users’ personalized preferences in the selection and recommendation process. The personalized preferences ensure that the non-functional attribute closely aligns with users’ satisfaction and thus accurately selects or recommend services to prospective users.

Impact

Including users’ personalized preferences in service selection and recommendation generally improves selection and recommendation accuracy. The models my team has developed are currently used to select airline services. We have also implemented one of our models to select additive manufacturing (3D printing) services in a cloud additive manufacturing framework. We have developed an intelligent survey question curation method. The new survey application improves respondent engagement and optimizes automatic questionnaire curation. Finally, with fake users in social media on the rise, we look to use our personalized preference-based methods. My team recently began this work and has derived some exciting results that we are going to share with the research community soon.

Other Research

Besides our work in service and cloud service selection and recommendation, my team has also undertaken research in additive manufacturing and its applications in the aerospace and automobile industry. Specifically, we have looked into part decomposition and tool path planning methods for metal additive manufacturing. We are currently working on a cloud additive manufacturing simulator, also for metal additive manufacturing processes, funded by the U.S. Department of Energy.

Researcher’s Background

I am an Assistant Professor of Computer Science at University of Massachusetts Boston (UMass Boston). I received my Ph.D. in 2015 from Missouri University of Science and Technology (Missouri S&T) in Computer Science with an emphasis in Software Design and Development. In addition, I also have an M.S. in Computer Science from Missouri S&T and a B.S. also in Computer Science from the Kwame Nkrumah University of Science and Technology, Ghana. Prior to joining UMass Boston, I worked as an Engineer in product innovation and developed software for Siemens, National Aeronautics and Space Administration, and BOEING. I have a patent and several publications in top conferences and journals in these research areas. My research spans services and cloud computing, machine and deep learning, software engineering, and additive manufacturing.