PROBLEM TO BE SOLVED: To quickly and accurately obtain data on a low-dimensional space from the examination value data of a new medical examinee by using data obtained by dimensionally compressing existing data.SOLUTION: A medical examination data processor includes: a medical examination database 111 for storing each examination value data of a predetermined examination item group for each medical examinee a compression database 113 for storing low-dimensional data mapped on a low-dimensional space for each medical examinee while maintaining inter-examination value data distance relation on a high-dimensional space stored in the medical examination database similar person extraction means S13 to S15 for comparing the examination value data of a candidate with the examination value data of each medical examinee stored in the medical examination database, and for extracting the predetermined number of medical examinees whose are close in distance and candidate data calculation means S17 for acquiring the low-dimensional data of each medical examinee extracted by the similar person extraction means from the compression database, and for calculating the low-dimensional data of the candidate on the basis thereof.COPYRIGHT: (C)2014,JPO&INPIT【課題】 既存のデータを次元圧縮したデータを用い、新規健診者の検査値データから低次元空間上のデータを速やかに精度よく得られるようにする。【解決手段】 所定の検査項目群の各検査値データを健診者毎に保持する健診データベース111,健診データベースが持つ高次元空間上の検査値データ間の距離関係を可能な限り保ちつつ低次元空間上に写像して得た低次元データを健診者毎に保持する圧縮データベース113,対象者の検査値データを健診データベースが持つ各健診者の検査値データと比較して距離が近い所定数の健診者を抽出する類似者抽出手段S13-S15,類似者抽出手段により抽出された各健診者の低次元データを圧縮データベースから取得してそれに基づいて対象者の低次元データを算出する対象者データ算出手段S17,を有する健診データ処理装置。【選択図】 図2
