Bài tập lý thuyết 2.2

23 views
Skip to first unread message

Cang Do

unread,
May 28, 2012, 11:43:31 AM5/28/12
to ch...@googlegroups.com
Give the sequence of S and G boundary sets computed by the CANDIDATE-ELIMINA- 
TION algorithm if it is given the sequence of training examples from Table 2.1 in 
reverse order. Although the final version space will be the same regardless of the 
sequence of examples (why?), the sets S and G computed at intermediate stages 
will, of course, depend on this sequence. Can you come up with ideas for ordering 
the training examples to minimize the sum of the sizes of these intermediate S and 
G sets for the H used in the EnjoySport example?

Cang Do

unread,
May 28, 2012, 9:56:03 PM5/28/12
to Forum thảo luận cho lớp cao học khóa 5
Bài này có lẽ phải nhờ các bạn góp ý để hiểu đề bài hơn.
Mình có thề tóm tắt ý mình hiểu tới như sau:

Thứ tự của tập dữ liệu training không quyết định kết quả cuối cùng là
S và G (giải thích ?)
Tuy nhiên các tập tính toán trung gian của chúng thì lại phụ thuộc vào
thứ tự đó.
Hãy đưa ra ý tưởng về quy tắt của thứ tự để cho "the sum of the sizes
of these intermediate S and G sets for the H" là nhỏ nhất ?

Nhờ giải thích thêm chỗ câu hỏi tiếng anh dùm !!!!
Reply all
Reply to author
Forward
0 new messages