CAJ | 학술논문

传统的数据库下挖掘方式以建立分类小区域为主，但是都以数据属性简单分类结果进行数据结构的划分，分类过程较为粗糙，存在冗余干扰，小区域内本身也存在着多余区域干扰。为此提出一种博弈最优区域数据结构下的数据库高效挖掘算法，首先通过子空间分解方法，在子空间中对数据进行分解分析，提取数据的深层次特征，再运用博弈的方法对层次数据进行细分，结构确定时，把建立区域的反应函数看成博弈过程，选取最优博弈结果，建立更加鲁棒的数据结构。采用一组拟定参数进行预测实验，实验结果说明，采用所提方法能够提高数据库的挖掘效率以及精确度。
전통적수거고하알굴방식이건립분류소구역위주，단시도이수거속성간단분류결과진행수거결구적화분，분류과정교위조조，존재용여간우，소구역내본신야존재착다여구역간우。위차제출일충박혁최우구역수거결구하적수거고고효알굴산법，수선통과자공간분해방법，재자공간중대수거진행분해분석，제취수거적심층차특정，재운용박혁적방법대층차수거진행세분，결구학정시，파건립구역적반응함수간성박혁과정，선취최우박혁결과，건립경가로봉적수거결구。채용일조의정삼수진행예측실험，실험결과설명，채용소제방법능구제고수거고적알굴효솔이급정학도。
Mining way under the traditional database to establish a classification of small area is given priority to, but with the division of data attribute classification results for the simple data structure, the classification process is relatively rough, there is redundancy interference, small area itself there is a spare area. Put forward a kind of game under the opti-mal regional data structure zone of database efficient mining algorithm, by using the method of subspace decomposition, the data in the subspace decomposition analysis, extract the data of deep characteristics, the article uses the method of game subdividing hierarchical data and structure, the response function as a game process, to establish a regional choose the optimal game as a result, a more robust data structure. Using worked out a set of parameters to predict the experi-ment, the experiment result shows that the proposed method can improve the efficiency and accuracy of database mining.