Title: Life Data Analysis of Binary Failure Data and Its Application

Report Abstract:

The life data analysis for failure/repair process has been well developed in statistical inference, where the calendar time is commonly used as the time axis. On one hand, when the cumulative number of successes in testing is taken as the time axis, the corresponding stochastic point process to represent the failure/repair process is reduced to a generalized Bernoulli process. Finkelstein (1983) dealt with a power-law type failure model and approximated it by the well-known power-law type non-homogeneous Poisson process (NHPPs). He also proposed the so-called continuous analog estimator (CAE) by regarding the binary data as a time data. Bhattacharyya and Ghosh (1991) proved asymptotic properties of the CAE.
In this talk, we show that the life data analysis can be essentially treated by using a simple Poisson-binomial distribution without the above Poisson approximation. An application in software testing is given as a typical example of life data analysis of the binary failure data. Through real data analysis with the actual software test data, we compare the Poisson-binomial model with the conventional NHPP model. It is emphasized that the recent development of high performance computing enables us to get the solution even for a large scaled problem.

Speaker: Tadashi Dohi, Professor

Date/Time: 15:00 – 17:00PM, October 12, 2019

Location: Room 715, CEM Building, Jiangjun Rd. Campus

Speaker Biography:

Tadashi Dohi received his BE, ME and Ph.D. degrees from Hiroshima University, Japan, in 1989, 1991 and 1995, respectively. In 1992 and 2000, he was a Visiting Researcher in University of British Columbia, Canada, and Duke University, USA, respectively, on the leave absent from Hiroshima University. Since 2002, he has been working as a Full Professor in the Department of Information Engineering, Graduate School of Engineering, Hiroshima University. His research areas include Reliability Engineering, Software Reliability, Dependable Computing, Performance Evaluation, Computer Security and Operations Research. He is a Regular Member of ORSJ, IEICE, IPSJ, REAJ, IEEE (Computer Society and Reliability Society). He published over 550 peer-reviewed papers, 40 book chapters, and edited 23 books/proceedings in the above research area. Dr. Dohi served as the General Chair of several international conferences including DASC 2019 and ORSJ Fall Meeting 2019. He has worked as a Program Committee Member in some major international premier conferences, and is an Associate Editor/Editorial Board Member in many international journals including IEEE Transactions on Reliability. He has been a President of REAJ and a Vice Dean of School of Informatics and Data Science, Hiroshima University since 2018.