第3期 程麟焰，等：基于模糊超网络的知识获取方法研究 ·487· 表6查准率 Table 6 Precision 序号 数据集 NaiveBayes KNN J48 SMO BP-KNN F-hypernetworks 1 BLOGGER 0.702 0.817 0.716 0.715 0.717 0.848 2 lymph 0.816 0.833 0.735 0.860 0.830 0.826 3 tae 0.518 0.657 0.518 0.580 0.487 0.568 4 flags 0.582 0.564 0.562 0.546 0.539 0.557 5 Glass 0.482 0.680 0.661 0.510 0.613 0.661 6 breast-cancer 0.716 0.681 0.662 0.677 0.714 0.712 7 Haberman 0.715 0.664 0.688 0.541 0.624 0.656 8 column_2C_weka 0.826 0.801 0.804 0.797 0.804 0.818 9 column 3C weka 0.826 0.789 0.831 0.772 0.759 0.822 10 ecoli 0.850 0.802 0.801 0.806 0.848 0.850 11 lonosphere 0.840 0.882 0.907 0.886 0.877 0.893 12 balance-scale 0.845 0.765 0.616 0.919 0.766 0.839 13 Pima diabetes 0.749 0.711 0.727 0.760 0.722 0.747 14 tic-tac-toe 0.687 0.985 0.842 0.984 0.985 0.845 15 car 0.846 0.921 0.907 0.935 0.921 0.904 16 平均值 0.733 0.770 0.732 0.753 0.747 0.770 表7查全率 Table 7 Recall 序号 数据集 NaiveBayes KNN J48 SMO BP-KNN F-hypernetworks 1 BLOGGER 0.720 0.820 0.730 0.730 0.730 0.850 2 lymph 0.818 0.831 0.736 0.858 0.838 0.824 3 tae 0.523 0.656 0.517 0.570 0.483 0.563 4 flags 0.562 0.567 0.572 0.557 0.588 0.593 5 Glass 0.472 0.682 0.668 0.561 0.636 0.673 6 breast-cancer 0.731 0.706 0.692 0.703 0.734 0.734 7 Haberman 0.748 0.680 0.735 0.735 0.706 0.729 8 column_2C_weka 0.781 0.787 0.803 0.800 0.803 0.810 9 column 3C_weka 0.829 0.784 0.832 0.755 0.758 0.819 10 ecoli 0.851 0.807 0.813 0.824 0.857 0.860 11 lonosphere 0.823 0.872 0.906 0.880 0.866 0.883 12 balance-scale 0.917 0.827 0.667 0.912 0.829 0.910 Pima diabetes 0.754 0.715 0.730 0.766 0.730 0.749 14 tic-tac-toe 0.701 0.984 0.843 0.983 0.984 0.847 15 car 0.853 0.922 0.907 0.933 0.922 0.905 16 平均值 0.739 0.776 0.743 0.771 0.764 0.783表 6 查准率 Table 6 Precision 序号 数据集 NaiveBayes KNN J48 SMO BP-KNN F-hypernetworks 1 BLOGGER 0.702 0.817 0.716 0.715 0.717 0.848 2 lymph 0.816 0.833 0.735 0.860 0.830 0.826 3 tae 0.518 0.657 0.518 0.580 0.487 0.568 4 flags 0.582 0.564 0.562 0.546 0.539 0.557 5 Glass 0.482 0.680 0.661 0.510 0.613 0.661 6 breast-cancer 0.716 0.681 0.662 0.677 0.714 0.712 7 Haberman 0.715 0.664 0.688 0.541 0.624 0.656 8 column_2C_weka 0.826 0.801 0.804 0.797 0.804 0.818 9 column_3C_weka 0.826 0.789 0.831 0.772 0.759 0.822 10 ecoli 0.850 0.802 0.801 0.806 0.848 0.850 11 Ionosphere 0.840 0.882 0.907 0.886 0.877 0.893 12 balance-scale 0.845 0.765 0.616 0.919 0.766 0.839 13 Pima_diabetes 0.749 0.711 0.727 0.760 0.722 0.747 14 tic-tac-toe 0.687 0.985 0.842 0.984 0.985 0.845 15 car 0.846 0.921 0.907 0.935 0.921 0.904 16 平均值 0.733 0.770 0.732 0.753 0.747 0.770 表 7 查全率 Table 7 Recall 序号 数据集 NaiveBayes KNN J48 SMO BP-KNN F-hypernetworks 1 BLOGGER 0.720 0.820 0.730 0.730 0.730 0.850 2 lymph 0.818 0.831 0.736 0.858 0.838 0.824 3 tae 0.523 0.656 0.517 0.570 0.483 0.563 4 flags 0.562 0.567 0.572 0.557 0.588 0.593 5 Glass 0.472 0.682 0.668 0.561 0.636 0.673 6 breast-cancer 0.731 0.706 0.692 0.703 0.734 0.734 7 Haberman 0.748 0.680 0.735 0.735 0.706 0.729 8 column_2C_weka 0.781 0.787 0.803 0.800 0.803 0.810 9 column_3C_weka 0.829 0.784 0.832 0.755 0.758 0.819 10 ecoli 0.851 0.807 0.813 0.824 0.857 0.860 11 Ionosphere 0.823 0.872 0.906 0.880 0.866 0.883 12 balance-scale 0.917 0.827 0.667 0.912 0.829 0.910 13 Pima_diabetes 0.754 0.715 0.730 0.766 0.730 0.749 14 tic-tac-toe 0.701 0.984 0.843 0.983 0.984 0.847 15 car 0.853 0.922 0.907 0.933 0.922 0.905 16 平均值 0.739 0.776 0.743 0.771 0.764 0.783 第 3 期 程麟焰，等：基于模糊超网络的知识获取方法研究 ·487·