| Project/Area Number |
17500179
|
|---|
| Research Category |
Grant-in-Aid for Scientific Research (C)
|
| Allocation Type | Single-year Grants |
|---|
| Section | 一般 |
|---|
| Research Field |
Statistical science
|
|---|
| Research Institution | Tokyo Institute of Technology |
|---|
Principal Investigator |
MIYAKAWA Masami Tokyo Institute of Technology, Graduate School, Department of Decision Science and Technology, Professor, 大学院社会理工学研究科, 教授 (90157595)
|
|---|
| Co-Investigator(Kenkyū-buntansha) |
TANAKA Kentaro Tokyo Institute of Technology, Graduate School, Department of Decision Science and Technology, Assistant Professor, 大学院社会理工学研究科, 助手 (00376948)
|
|---|
| Project Period (FY) |
2005 – 2006
|
| Project Status |
Completed (Fiscal Year 2006)
|
| Budget Amount *help |
¥3,000,000 (Direct Cost: ¥3,000,000)
Fiscal Year 2006: ¥1,300,000 (Direct Cost: ¥1,300,000)
Fiscal Year 2005: ¥1,700,000 (Direct Cost: ¥1,700,000)
|
|---|
| Keywords | probability theory / mathematical engineering / statistical science / quality control system / statistical modeling |
|---|
| Research Abstract |
This research studies several identification problems in statistical models.
1)Intervention not to variables but arrows in the causal diagram is formulated through the conditional intervention, and the identifiability criterion is shown.
2)In the competing risks with covariates, a new identifiable assumption which can reduce high dimensional covariates is obtained.
3)Strong consistency of the maximum likelihood estimator for finite mixtures of location-scale distributions is proved in the case where the scale parameters of the component distributions are exponentially small.
4)The way to assign six-level factors to the L36 orthogonal array is given. Three six-1evel factors with eight two-level factors can be assigned orthogonally. Under the least confounding, six six-level factors are assigned and all main effects are estimable. The way of assignment is shown by nearly orthogonal arrays.
5)The error rate of a future observation from a normal group is investigated in the Mahalanobis-Taguchi system. It is shown that by setting a nominal error based on not chi-square distribution but F distribution, the expected error rate coincides with the nominal value and the conditional error rates tend to be stable.
|
