决策树算法在临床诊断中应用的研究.pdfVIP

摘 要 数据挖掘是信息处理领域的一项重要课题，它融合了数据库、人工智能、机器 学习、统计学等多个领域的理论和技术。而决策树方法以其速度快、精度高、生成 的模型简单易懂以及容易转化成分类规则等优点，在数据挖掘领域被广泛地研究和 应用。本文采用决策树算法作为核心的数据挖掘算法进行临床诊断数据挖掘。在生 成决策树时，可能使得产生的决策树非常庞大，对应的分类规则也难以理解，需要 对生成树进行剪枝。因此剪枝算法在决策树算法 中占有很重要的地位。针对数据挖 掘在临床诊断中的应用，通过对经典决策树剪枝算法的分析，提出了一种基于多策 略思想的剪枝算法。该算法从用户需求的出发，根据不同的数据挖掘集接受对决策 树模型描述性的参数，最终得到理想的决策树模型。实验结果表明，该算法能够很 好地平衡剪枝算法的精确性和复杂性，满足不同医学诊断应用的需求，保证了对不 同数据挖掘集取得更好的适应性。 关键词：临床诊断；决策树；剪枝算法；多策略 Abstract Data mining is the most signifieant field of information teehnology, which is combined with many theory and technology such as database, artificial intelligence, machine learning and statisties. As one of the most widely used classification algoritms of data mining, decision tree algorithm is characterized by high accuracy, simplicity, high efficiency, and can be translated to rules easily. This paper uses decision tree algorithm as the data mining algorithm for mining medical diagnosis data. The decision tree may be very huge when the decision tree was first generated, and the corresponding rules will be very hard to understand. So it needs pruning algorithm to reduce the scale of the decision tree. So pruning algorithm is very important in decision tree algorithms. This paper focuses on the application of data mining in medical diagnosis, and analyses the advantage and disadvantage of classic decision tree pruning algorithms, presents a new pruning algorithm based on multi-strategy. This algorithm considers user’s needs, and can accepts descriptive parameters of decision trees on different data mining set. Finally, it gets the ideal decision tree model. The results of experiments show that this algorithm can balances the precision and complexity better, and me