摘 要：In this talk, I will present some examples of current work at IBM Research that leverages advanced statistics to address services-related business issues. In particular, I will present a new technique for model-based clustering of noisy, high-dimensional sequences. The developed technique was motivated by a need to create a resource requirement-based project taxonomy for workforce forecasting and optimization by mining records of observed resource utilization for a set of historical projects. For modeling sequences, we use explicit duration hidden semi-Markov models (HsMM) capable of handling sparse and noisy Dirichlet-distributed observations and having a left-to-right topology representing the project lifecycle. We introduce an efficient and optimal sequence decoding algorithm for this HsMM topology and demonstrate its effectiveness both theoretically and experimentally. EM-like algorithms for sequence modeling and clustering are employed.

作者简介：Bonnie Rayis currently Program Manager, Data and Analytics, IBM China Research Lab. Her area of expertise is applied statistics, with particular interest in the use of statistics for business analytics. Since joining IBM, she has played key roles in developing customer targeting models and tools for IBM’s outsourcing businesses, resource demand forecasting methods for workforce management processes, and methods and tools for automated risk assessment in software development. She has also been responsible for managing the relationship between IBM Research and IBM’s technology interests in the Consumer Products industry. Prior to joining IBM, she was a tenured faculty member in the Mathematics department at the New Jersey Institute of Technology, and has held visiting appointments at Stanford University, the University of Texas, and the Los Alamos National Lab. She holds a Ph. D. in Statistics from Columbia University and a B.S. in Mathematics from Baylor University.