分类模式挖掘在属性预测中的应用

Application of Classification Mining in Attribution Prediction

数据挖据是一种处理海量数据的技术。分类挖掘是数据挖掘的重要方法。决策树算法能有效在训练数据集上建立数据属性和类别的映射。利用决策树算法建立目标数据库分类器,对数据对象的缺失属性预测。针对分类预测模型对单个目标多个预测类别的现象,提出一种单个目标预测结果的综合分析算法,处理结果得到单个目标的预测类别及其可信度。分类结果可用于空缺或错误字段补全或校正。

Data mining is a technology for processing mass data. Classification mining is an important method of data mining. Decision tree algorithm generates the map between attributes and class labels in training data sets. A classifier of object database is constructed based on decision tree algorithm and is used to predict the null fields. A method of analyzing multi-prediction results is proposed, and the final prediction result and confidence for a single target is obtained. Classification results are used in patching the null and emendation.