Đang chuẩn bị liên kết để tải về tài liệu:
KNOWLEDGE-BASED SOFTWARE ENGINEERING phần 10

Tham khảo tài liệu 'knowledge-based software engineering phần 10', công nghệ thông tin, kỹ thuật lập trình phục vụ nhu cầu học tập, nghiên cứu và làm việc hiệu quả | M. Komori et al. A New Feature Selection Method 295 Table 2 The accuracy respectively performed by three feature selection methods No Dataset No feature selection method Filter method Wrapper method Seed method 1 breast 95.14 95.14 95.71 o 95.56 2 crx 85.94 - 85.65 86.96 o - 85.8 3 Hayes-Roth 92.86 o 92.86 o - 71.43 92.86 o 4 labor-neg 82.35 o 82.35 o - 76.47 82.35 o 5 pima 74.08 74.08 75.0 o 75.0 o 6 sick 98.77 98.77 98.77 99.18 o 7 audiology 84.62 - 81.62 92.31 o 92.31 o 8 chess 99.53 o - 99.47 - 97.87 99.53 o 9 glass 65.89 65.89 71.96 72.90 o 10 lung-cancer 81.25 81.25 84.38 o 84.38 o 11 wine 94.94 94.94 97.50 o 97.2 12 monkl 75.69 88.89 o 88.89 o 88.89 o 13 monk2 65.05 - 62.50 67.13 o 67.13 o 14 monk 3 97.22 o 97.22 0 97.22 o 97.22 o Ave 85.24 85.76 85.83 87.88 takes the first place from the viewpoint of accuracy and the second place from point of computational costs. As future work we will have more theoretical analysis on our seed method and apply it to different kinds of other data sets. Table 3 The computational costs respectively taken by three feature selection methods No Dataset Filter method Wrapper method Seed method 1 breast 37 26 44 2 crx 43 782 100 3 Hayes-Roth 1 1 4 4 labor-neg 1 16 64 5 pima 29 304 20 6 sick 126 1708 352 7 audiology 2 183 752 8 chess 1753 9060 668 9 glass 4 411 28 10 lung-cancer 1 44 216 11 wine 3 176 68 12 monkl 1 3 20 13 monk2 1 2 12 14 monk3 1 3 16 Ave 143.07 992.14 168.86 296 M. Komori et al. A New Feature Selection Method Table 4 The number of features respectively selected by three feature selection methods No Dataset all the features the features with no feature selection the features with filter method the features with wrapper method the features with seed method 1 breast 10 7 10 4 5 7 crx 15 9 14 8 8 3 Hayes-Roth 4 3 3 1 3 4 labor-neg 16 2 13 1 5 pima 8 6 8 6 7 6 sick 29 10 25 11 24 7 audiology 69 14 42 9 9 8 chess 36 22 28 12 24 9 glass 10 9 9 5 7 10 lung-cancer 56 2 31 2 2 11 Wine 13 3 13 4 6 12 monkl 6 5 3 3 4 13 .