→ LiamIssac: 先把ABC定義好 矩陣照著打就好了 09/26 18:30
推 profyang: A是純量的話 用[diag(A*ones(4,1)),zeros(4,2)]這樣去弄 09/26 18:47
→ profyang: 不對等等 用[A*eye(4),zeros(4,2)]就可了 09/26 18:48
→ Ecampus: !! 感謝 初學 我程式SENSE太差ㄏㄏ 09/26 18:53
→ YoursEver: toeplitz matrix; just call "toeplitz" please. 09/27 08:40
→ YoursEver: or: gg=[1 2 3]; hh=eye(10); ff=conv2(hh,gg,'same') 09/27 08:46
→ YoursEver: then, modify ff to obtain the matrix you need. 09/27 08:46
→ Ecampus: 我後來把A B C各自寫成對角線矩陣 再把B和C做移位 09/29 01:49
→ Ecampus: 然後相加XD... 09/29 01:49
→ YoursEver: 要看你的用途,如果你是要把1D convolution kernel套在 09/29 08:02
→ YoursEver: matrix form寫成數學式來表示,計算並藉此推導closed 09/29 08:04
→ YoursEver: form的話,最後的計算免不了要使用toeplitz matrix, 09/29 08:04
→ YoursEver: 那麼,靠系統化一點的方式來產生大矩陣才會更恰當. 09/29 08:06
→ Ecampus: 3Q 我再練習一下 10/05 23:39