<< Back to Portfolio << Previous Next >>
-----------------------------------------------------------------------------------------------------------------------------------------------------------------------
Research Projects
How are you performing? A study of employee performance review management process in organizations
Employee performance reviews: while often a requirement in many organizations, performance appraisals can be a frustrating time for managers, employees and human resources professionals alike. In recent years, a growing number of automated software systems have become available to organizations attempting to solve the CSCW challenges of providing employee feedback. We interviewed eight individuals at four different organizations to gain insight into their existing processes and systems. We synthesized the discoveries made during the interviews and the evaluations of the automated systems. The findings from the interviews were quite similar, despite the diversity of organizations and interviewees. While progress has been made in the field of automated performance review systems, the socio-technical gap continues to challenge the successful implementation and ongoing use of these products.
Click here to see the full paper
Guide: Prof. Mark Ackerman, University of Michigan
Team members: Emily Luke, Megan Schwarz, Melissa Cox and Suyog Deshpande
Affiliation: School of Information, University of Michigan, Ann Arbor - 2011
Gift Recommender System for Amazon
This paper suggests a gift recommender system for Amazon to provide timely and appropriate suggestions that have higher probability of getting converted into an actual purchase. The users can be a lot more confused while making a purchase for others as compared to while making a purchase for themselves. So, in such a scenario a system that can both organize important events and dates and at the same time provide good recommendations suiting the event/occasion can be very helpful both to the users as well as sellers. The proposed recommender is an improvement over the existing gift recommender of Amazon.
Click here to access the report.
Guide: Prof. Rahul Sami, University of Michigan
Affiliation: School of Information, University of Michigan, Ann Arbor - 2011
Reputation System for couch surfing portals and websites
The concept of couch surfing is seeing the world from a stranger’s couch/sofa. The website “Couchsurfing.com”, which is the market leader in this domain. Since all the contributions from users on this website are voluntary, great care needs to be taken while implementing any engagement or safeguarding scheme, as it might discourage the users and may be even lead users to withdraw from the system. This paper talks about the potential of reputation system in this domain with discussion about the existing reputation techniques employed by couchsurfing.com, their shortcomings and proposes two reputation techniques for better engagement and safeguarding.
Click here to access the report.
Guide: Prof. Rahul Sami, University of Michigan
Affiliation: School of Information, University of Michigan, Ann Arbor - 2011
Researching parameters supporting successful collaboration in autonomous systems
The main focus of this project was to analyze the impact of providing task related information to the task doers during different stages of collaborative interaction. Relevant user testing was carried out to verify the formulated hypothesis.
This entire experiment was carried out as a part of a PhD study researching “The inputs needed for Autonomous systems to collaborate”. This was a Defense Technology Center funded project.
Guide: Prof. Peter Johnson, University of Bath, UK
Team members: Rachid (Doctoral student)
Affiliation: HCI group, Computer Science Department, University of Bath, UK - 2007
Note: I cannot reveal the documentation and report due to copyright agreement.
-----------------------------------------------------------------------------------------------------------------------------------------------------------------------
<< Back to Portfolio << Previous Next >>