| Project/Area Number |
15300072
|
|---|
| Research Category |
Grant-in-Aid for Scientific Research (B)
|
| Allocation Type | Single-year Grants |
|---|
| Section | 一般 |
|---|
| Research Field |
Sensitivity informatics/Soft computing
|
|---|
| Research Institution | The University of Tokushima |
|---|
Principal Investigator |
AOE Jun-ichi The University of Tokushima, Institute of Technology and Science, Professor, 大学院ソシオテクノサイエンス研究部, 教授 (90108853)
|
|---|
| Co-Investigator(Kenkyū-buntansha) |
KITA Kenji The University of Tokushima, Center for advanced information technology, Professor, 高度情報化基盤センター, 教授 (10243734)
SHISHIBORI Masami The University of Tokushima, Institute of Technology and Science, Assistant professor, 大学院ソシオテクノサイエンス研究部, 助教授 (50274262)
FUKETA Masao The University of Tokushima, Institute of Technology and Science, Assistant professor, 大学院ソシオテクノサイエンス研究部, 助教授 (10304552)
MORITA Kazuhiro The University of Tokushima, Institute of Technology and Science, Assistant professor, 大学院ソシオテクノサイエンス研究部, 講師 (20325252)
|
|---|
| Project Period (FY) |
2003 – 2006
|
| Project Status |
Completed (Fiscal Year 2006)
|
| Budget Amount *help |
¥13,800,000 (Direct Cost: ¥13,800,000)
Fiscal Year 2006: ¥3,500,000 (Direct Cost: ¥3,500,000)
Fiscal Year 2005: ¥3,000,000 (Direct Cost: ¥3,000,000)
Fiscal Year 2004: ¥3,500,000 (Direct Cost: ¥3,500,000)
Fiscal Year 2003: ¥3,800,000 (Direct Cost: ¥3,800,000)
|
|---|
| Keywords | Semantic collocation / emotion expression / Retrieval for sensibility information / speech dialogue / dialogue understanding / Processing of natural language / 自然言語理解 |
|---|
| Research Abstract |
In 2003
1) Research for construction of knowledge about emotion collocation
a) Collection and analysis for sensibilities by language which express one's emotions
Categorized information about emotion words into 5, and confirmed reliability of the emotion information by cooperating with at least five people.
b) Registration of semantic attribute into a dictionary database for analysis which analyzed Japanese
Comprised emotion information which was analyzed and categorized into a dictionary database for analysis
2) Research for collection and analysis for emotion expressions with relation of semantic collocation, and high-speed matching technology
a) Research for collection, analysis and construction of emotion collocation
Collected and analyzed emotion information for semantic collocation by modality, and comprised knowledge of emotion rules.
b) Experiment for high-speed matching algorithm of semantic collocation
Built semantic collocations into an analysis engine for Japanese, and conducted an e
… More
valuation experiment for algorithm which judged collocation and was able to extract emotion information at high speed. Changed abstracted concepts for kana-kanji conversion, conducted a experiment for matching at high-speed and improved the method.
After 2004
1) Research for analysis algorithm for Japanese which prioritized meanings and evaluated relations of emotion collocation at the same time
a) Research for analysis for Japanese which was able to evaluate constructions and meanings at the same time.
Studied the technology which made evaluation technology of semantic collocation for kana-kanji conversion understand emotion knowledge.
b) Analysis and evaluation for results of emotion understanding, and improvement
As there are various emotion expressions, evaluation experiment by cooperating with many people needed to be conducted in order to determine the strength of emotion depending on positions of speakers and listeners. The accuracy of emotion
understanding was improved by this experiment.
2) Research for retrieval of emotion information
a) Research for filtering emotion information by sensibility retrieval
Studied analysis of quantitative number attribute related to sensibility.
b) Evaluation experiment and improvement for information retrieval by sensibility keywords
Applied sensibility keywords to contents and electronic mail database, evaluated characteristics and effectiveness of keywords and improved the suggested method. Less
|
