An Integrative Personal Character Model and Its Modeling with Scenario-based Fusion

• Material Type: 博士論文
• Title: An Integrative Personal Character Model and Its Modeling with Scenario-based Fusion
• Author/Editor: GUO, Ao
• Author Heading: GUO, Ao
• Publication Date: 2021-03-24
• Publication Date (W3CDTF): 2021-03-24
• Degree grantor/type: 法政大学 (Hosei University)
• Date Granted: 2021-03-24
• Date Granted (W3CDTF): 2021-03-24
• Dissertation Number: 甲第514号
• Degree Type: 博士(理学)
• Conferring No. (Dissertation): 甲第514号
• Text Language Code: eng
• Subject Heading: Personal Character; Characteristic; Personality; Affect; Behavior; Cognition; Multi-Scenario; Sub-Scenario; Scenario-based Fusion
• Target Audience: 一般
• Note (General): type:Thesis; An individual-like intelligent artifact is a man-made physical or digital thing that resembles a real individual. Such an intelligent artifact is capable of helping achieve numerous fantastic applications, including better-personalized services, the succession of individual’s work style, and can be extended to individual life. As the AI with personal character, the individual-like intelligent artifact can be implemented by a physical robot, a digital avatar or an invisible software. Regardless of the form of individual-like intelligent artifacts are implemented, the individual-like intelligent artifacts would have the same personal characters with its corresponding humans. Personal character refers to all the personal characteristics that characterize an individual. Two fundamental problems with the research on personal character are identified: (1) there is no comprehensive and structuralized description for personal character and (2) it remains unclear about how to build the personal character model accurately and comprehensively. Correspondingly, this research consists of two parts: personal character model and personal character modeling. In the first part, an integrative personal character model is proposed by considering differential psychology and personality psychology. It covers macro characteristics of personality (P), micro characteristics in affect (A), behavior (B) and cognition (C), and relational characteristics among characteristics (R). The proposed model is known as the ABC-P-R model. A stepwise computation process and RDF-based representation scheme for the data collected from twenty participants are presented for the construction of personal character model. The proposed integrative personal character model is verified to be computable, with provided existence and rationality of relational characteristics. As personality is merely one aspect of characteristics in the personal character model, the modeling is built by using monomodal classification and multimodal fusion of personality computing. However, these common methods are to compute one or a few personal characteristics by using the data collected from a uni-scenario or an experiment. Such approach is insufficient to achieve a comprehensive and accurate classification of the whole characteristics of personal character. A scenario-based fusion for personal character modeling, consisting of sub-scenario fusion and multi-scenario fusion, is proposed to address such problem. The state corresponding to individual’s experiment during data collection is defined as a scenario, whereas a sub-scenario is a part of the whole scenario. The sub-scenario fusion is conducted to split the data collected from a single scenario into several sub-scenario data and then to fuse these sub-scenario data. Multi-scenario fusion is performed to fuse the data collected from different scenarios or experiments. The fusion frameworks corresponding to these two scenario-based fusions are presented, in which both sub-scenario and multi-scenario fusion can be implemented by feature fusion and classifier fusion, and multi-scenario fusion can be additionally implemented by using multi-scenario incremental fusion. Five datasets are used for concrete computations and evaluation in personal character modeling, including Essay dataset, YouTube dataset, Pan15 dataset, MyPersonality dataset, and Physiological dataset. Scenario-based fusion method is evaluated by using these five datasets, and the fusion method proved to be effective to improve the performance in the given personal data and the collected scenario of data.
• DOI: 10.15002/00024121; https://doi.org/10.15002/00024121
• Persistent ID (NDL): info:ndljp/pid/11692112; https://dl.ndl.go.jp/pid/11692112
• Collection: 障害者向け資料
• Collection (Materials For Handicapped People:1): テキストデータ
• Collection (particular): 国立国会図書館デジタルコレクション > デジタル化資料 > 博士論文; https://dl.ndl.go.jp/collections/A00014
• Acquisition Basis: 博士論文（自動収集）
• Date Accepted (W3CDTF): 2021-07-05T22:24:43+09:00
• Date Created (W3CDTF): 2021-06-17
• Format (IMT): PDF; application/pdf
• Access Restrictions: 国立国会図書館内限定公開
• Service for the Digitized Contents Transmission Service: 図書館・個人送信対象外
• Availability of remote photoduplication service: 可
• Periodical Title (URI): https://doi.org/10.15002/00024121; http://hdl.handle.net/10114/00024121
• Data Provider (Database): 国立国会図書館 : 国立国会図書館デジタルコレクション; https://dl.ndl.go.jp