Building a Recommendation System with Limited Data
Share this Session:
  Cirrus Shakeri   Cirrus Shakeri
Senior Director
SAP Labs


Thursday, August 21, 2014
10:30 AM - 11:00 AM

Level:  Technical - Intermediate

Most recommendation systems in use today, especially in the consumer Internet domain (e.g., Amazon or Netflix), rely on the availability of large amounts of data and the use of machine learning algorithms. In the enterprise applications, however, smaller volumes of data are available (especially during the initial phases of deploy the application). This limits the effectiveness of machine learning algorithms in building recommendations systems. The data privacy requirements in the enterprises further limits the available data and thus the use of machine learning. In this session we report on a case study where we’ve combined machine learning algorithms with the graph-based approach to address the problem of limited data in building a recommendation system. The attendees will learn about the advantages of this hybrid approach and its challenges in the context of a specific enterprise application.

Cirrus Shakeri works in the HANA Platform Strategic Projects at SAP Labs in California focusing on the applications of AI and semantic technologies in enterprise software. He has over fifteen years of experience in technology innovation and new product development including applied research, software development lifecycle, product management, customer co-innovation, enterprise-wide solution deployment, and closing the cycle by taking the opportunities for developing new solutions back to R&D. He received a Ph.D. in AI and Engineering from Worcester Polytechnic Institute in Massachusetts in 1998. He is a member of the IEEE Computer Society.

Close Window