{"created":"2023-06-20T13:20:42.801168+00:00","id":748,"links":{},"metadata":{"_buckets":{"deposit":"ff2f216d-4235-479d-b0b9-71f1e348f93d"},"_deposit":{"created_by":1,"id":"748","owners":[1],"pid":{"revision_id":0,"type":"depid","value":"748"},"status":"published"},"_oai":{"id":"oai:ir.soken.ac.jp:00000748","sets":["2:429:17"]},"author_link":["9065","9066","9067"],"item_1_creator_2":{"attribute_name":"著者名","attribute_type":"creator","attribute_value_mlt":[{"creatorNames":[{"creatorName":"三分一, 史和"}],"nameIdentifiers":[{}]}]},"item_1_creator_3":{"attribute_name":"フリガナ","attribute_type":"creator","attribute_value_mlt":[{"creatorNames":[{"creatorName":"ミワケイチ, フミカズ"}],"nameIdentifiers":[{}]}]},"item_1_date_granted_11":{"attribute_name":"学位授与年月日","attribute_value_mlt":[{"subitem_dategranted":"2001-03-23"}]},"item_1_degree_grantor_5":{"attribute_name":"学位授与機関","attribute_value_mlt":[{"subitem_degreegrantor":[{"subitem_degreegrantor_name":"総合研究大学院大学"}]}]},"item_1_degree_name_6":{"attribute_name":"学位名","attribute_value_mlt":[{"subitem_degreename":"博士（学術）"}]},"item_1_description_12":{"attribute_name":"要旨","attribute_value_mlt":[{"subitem_description":"　This thesis presents new strategies for analyzing the highly non-linear and complex time series that arise from recording the electrical signals of the brain (EEG) of epileptic patients.  These new strategies are necessary because, to date, there is no satisfactory theoretical or empirical model for this type of time series.  The availability of simple and interpretable time series models for epileptic data has been recognized as a prerequisite for improving the diagnosis and treatment of this disease.  As will be shown in the body of this work a judicious use of available nonparametric and semi-parametric autoregressive time series models will reveal the structure of his type of data.  Studies in this thesis lie on the boundaries of several fields:  applied statistics, nonlinear dynamics and brain science.  Therefore our goal was to use existing techniques to uncover hidden information about the up to know not well understood and less well modeled phenomenon of epileptic seizures.<br />　A particular segment of one of the channels of epileptic BEG so called \"Spike and Wave (SW)\" was analyzed in this work.  SW appears during \"absence seizure\" in which patient had stared fixedly into space with loss of contact with the external world.  Because of complexity and highly non-linearity, SW have attracted the attention of researchers in many field, psychiatrists, satiations, physicists and so on.<br /><br />　At first the comparison of the fully and popular non-parametric models, Nadaraya-Watson(NW) kernel estimator, Local Linear Polynomial Regression(LLPR) and Support Vector Machine(SVM), for SW activity was investigated.  This thesis describes, to our knowledge, the first application of Support Vector Machine Regression to ERG data.  It is shown that automatic methods for the selection of tuning parameters are not optimal for finding mimetic models for SW and Local Linear Polynomial Regression does not produce mimetic simulations either in the noise free or the stochastic case.  The performance of the simplest of all the models, the NW regression, is the most stable to stochastic simulations.<br /><br />　And next, the fully nonparametric model for the SW is successively simplified by developing of structured nonparametric regression models (Partial Linear Model, Additive model, Single Index Model and these combined models).  A number of conclusions seem to be warranted.<br />　This work contains, to our knowledge, the first use of partial linear, additive and single index models to EEG data.  The criteria for evaluation are NOT one step ahead prediction only but the much more stringent criterion of being mimetic.<br /><br />　And Partial Linear Additive Model was newly suggested in this work.  It is a combined model of Partial Linear Model and Additive Model and shown for the first time that SW activity may be separated into the sum of linear and non-linear components, and moreover this non-linear component can be split into some additive non-parametric component by use of this new Model.  The separate observation of the linear and nonlinear components suggests new interpretation in terms of nonlinear dynamics.<br />　Also it is a first time that an extremely simple model may describe the genesis of the SW models, consisting of a linear part and some bi-variate no-linear functions.  Further parametric modeling may be based upon these results.<br />　Single index models reveal for the first time static non-linearities of the sigmoid type that array suggest lines of physiological modeling.<br /><br />　 Through this thesis, these results will contribute to approach to physiological and physical neural mass model for SW.  And also it will be useful and helpful suggestions and motivations for the researchers who are studying epileptic EEG.","subitem_description_type":"Other"}]},"item_1_description_7":{"attribute_name":"学位記番号","attribute_value_mlt":[{"subitem_description":"総研大甲第501号","subitem_description_type":"Other"}]},"item_1_select_14":{"attribute_name":"所蔵","attribute_value_mlt":[{"subitem_select_item":"有"}]},"item_1_select_8":{"attribute_name":"研究科","attribute_value_mlt":[{"subitem_select_item":"数物科学研究科"}]},"item_1_select_9":{"attribute_name":"専攻","attribute_value_mlt":[{"subitem_select_item":"15 統計科学専攻"}]},"item_1_text_10":{"attribute_name":"学位授与年度","attribute_value_mlt":[{"subitem_text_value":"2000"}]},"item_creator":{"attribute_name":"著者","attribute_type":"creator","attribute_value_mlt":[{"creatorNames":[{"creatorName":"MIWAKEICHI, Fumikazu","creatorNameLang":"en"}],"nameIdentifiers":[{}]}]},"item_files":{"attribute_name":"ファイル情報","attribute_type":"file","attribute_value_mlt":[{"accessrole":"open_date","date":[{"dateType":"Available","dateValue":"2016-02-17"}],"displaytype":"simple","filename":"甲501_要旨.pdf","filesize":[{"value":"292.0 kB"}],"format":"application/pdf","licensetype":"license_11","mimetype":"application/pdf","url":{"label":"要旨・審査要旨 / Abstract, Screening Result","url":"https://ir.soken.ac.jp/record/748/files/甲501_要旨.pdf"},"version_id":"0b3c6288-a75a-469a-b91c-e9c011e9e2ce"}]},"item_language":{"attribute_name":"言語","attribute_value_mlt":[{"subitem_language":"eng"}]},"item_resource_type":{"attribute_name":"資源タイプ","attribute_value_mlt":[{"resourcetype":"thesis","resourceuri":"http://purl.org/coar/resource_type/c_46ec"}]},"item_title":"Characterization of Spike ＆ Wave Signals inEpileptic EEG: A non-linear non-parametric timeseries approach","item_titles":{"attribute_name":"タイトル","attribute_value_mlt":[{"subitem_title":"Characterization of Spike ＆ Wave Signals inEpileptic EEG: A non-linear non-parametric timeseries approach"},{"subitem_title":"Characterization of Spike ＆ Wave Signals inEpileptic EEG: A non-linear non-parametric timeseries approach","subitem_title_language":"en"}]},"item_type_id":"1","owner":"1","path":["17"],"pubdate":{"attribute_name":"公開日","attribute_value":"2010-02-22"},"publish_date":"2010-02-22","publish_status":"0","recid":"748","relation_version_is_last":true,"title":["Characterization of Spike ＆ Wave Signals inEpileptic EEG: A non-linear non-parametric timeseries approach"],"weko_creator_id":"1","weko_shared_id":1},"updated":"2023-06-20T14:50:05.524933+00:00"}