基于支持向量机的不平衡数据分类算法研究-计算机技术专业论文.docxVIP

大摘要 大摘要 万方数据 万方数据 提高分类性能。 综上所述，本文对支持向量机进行了改进，使其能更好地用于解决类不平 衡问题。然而，本研究仍有许多待改进和提高之处，如所提算法尽管可取得较 好的分类效果，但时空复杂度仍然较高，这也有待在未来的工作中做出改进。 另外，本文涉及的集成分类算法均是以 Bagging 为框架开发的，故在未来的工 作中，也将进一步考虑现有算法与 AdaBoost 集成学习框架相结合的可行性。同 时，本文算法只在少量数据集上进行了验证，在未来的工作中，我们也希望能 在各种应用领域的真实数据集上对本文算法进行验证，以推广其应用的范围。 关键词：支持向量机；类不平衡学习；集成学习；分类；DNA 微阵列数据 Ab Abstract The Research of Imbalanced Data Classification Algorithm Based on Support Vector Machine Abstract In the era of information explosion, the large number of data has aroused peoples attention, thus it needs people to find their own regular patterns and to make full use of them. Classif ication problem are one of the most frequently encountered problems in data processing. It has become an important research content of machine learning. Compared with the traditional classification methods, support vector machine has several merits as follows: high generalization ability, absence of local minima and adaptation for high-dimens ion and small sample data, which can better solve the problems listed as follows: over-learning, dimens ion disaster and local minima, thus we give priority to support vector machine (SVM) in this thesis. The main idea of Support Vector Machine (SVM) is that it makes the training set be mapped to a high dimension space by using one kernel function. Previous studies have shown that support vector machine (SVM) has the better classification effect on the balanced datasets, but for the imbalanced datasets, it is difficult to get excellent results. The main reason is that classification hyperplane of Support Vector Machine (SVM) is only decided by a small number of support vectors . When support vector machine (SVM) classifier is used to deal with class imbalance problem, its prediction is biased. The more the number of samples, the smaller its classification error is, and vice versa. So the major research work of this thesis is to study how to address class imbalanced problem by SVM. The main research results include the fo