| Project/Area Number |
15H02750
|
|---|
| Research Category |
Grant-in-Aid for Scientific Research (B)
|
| Allocation Type | Single-year Grants |
|---|
| Section | 一般 |
|---|
| Research Field |
Intelligent informatics
|
|---|
| Research Institution | Shimane University |
|---|
Principal Investigator |
|
|---|
| Co-Investigator(Kenkyū-buntansha) |
平野 章二 島根大学, 医学部, 准教授 (60333506)
河村 敏彦 島根大学, 医学部, 准教授 (70435494)
|
|---|
| Research Collaborator |
IWATA Haruko
KONDOH Masafumi
TANAKA Yoichiro
GOBARA Atsushi
KIMURA Tomohiro
|
|---|
| Project Period (FY) |
2015-04-01 – 2018-03-31
|
| Project Status |
Completed (Fiscal Year 2017)
|
| Budget Amount *help |
¥13,780,000 (Direct Cost: ¥10,600,000、Indirect Cost: ¥3,180,000)
Fiscal Year 2017: ¥3,900,000 (Direct Cost: ¥3,000,000、Indirect Cost: ¥900,000)
Fiscal Year 2016: ¥4,290,000 (Direct Cost: ¥3,300,000、Indirect Cost: ¥990,000)
Fiscal Year 2015: ¥5,590,000 (Direct Cost: ¥4,300,000、Indirect Cost: ¥1,290,000)
|
|---|
| Keywords | データマイニング / アクティブマイニング / 時系列マイニング / クラスタリング / クリニカルパス / 病院情報システム / クラスタ分析 / 類似度 / 人工知能 / 機械学習 |
|---|
| Outline of Final Research Achievements |
This project proposes automated construction of clinical pathways from the service logs stored in a hospital information system (HIS) in the following five steps. First, service logs and discharge summaries are extracted from HIS. Then, EM clustering applies all the cases with the same DPC code and decomposes them into subgroups. Third, to each subgroup, construction of clinical pathways based on dual clustering is applied. Fourth, classifiers are constructed by using combination of correspondence analysis and classification learning methods, such as decision tree, random forest, SVM and neural networks. Then,finally classifiers and clinical pathways are linked together,The proposed method was evaluated on HIS in Shimane University Hospital, whose results show that the proposed method generate more accuratebclinical pathways.
|
