BBS: Inland Empire Archive
Date: 05-02-92 (19:01) Number: 136
From: PAUL LEONARD Refer#: NONE
To: RICHARD VANNOY Recvd: NO
Subj: Matrix Conf: (2) Quik_Bas
On or about <May 01 19:40>, Richard Vannoy (1:105/314) scribbled:
RV> MatrixA MatrixB MatrixC
RV> 4 5 6 3 2 6 5 4 3 0 E M P T Y
RV> 3 2 3 4 5 5 4 3 3 3
RV> 2 1 0 6 9 5 2 1 0 4
RV> so, something like...
RV> MAT(C) = MAT(A) * MAT(B)
(nit-picking mode on)
My linear algebra's a bit rusty, but i don't think you can multiply the
matrices in your example. If A is rA x cA (rA rows by cA
columns), and B is rB x cB, then cA must equal rB in order
to multiply the two matrices. This is because the
elements of C are formed by multiplying the rows of A by
the columns of B, with C being rA x cB (i learned this by
thinking of the two equal dimensions needing to "canceling
out"). This is why, in general, AB <> BA. AB is rA x cB
and BA is rB x cA.
(nit-picking mode off)
(For Randy Baer)
I was once put through the matrix multiplication wringer by
a sadistic Fortran professor - he reacted rather badly to
a negative student evaluation of his class and vengefully
assigned a nasty stew of matrix manipulation (without MAT
functions), statistical analysis and formatting. Anyway,
the bare bones of the matrix multiplication are...
DIM mat.a%(rA%,cA%)
DIM mat.b%(rB%,cB%)
DIM mat.c%(rA%,cB%)
FOR i%=1 TO rA%
FOR j%=1 TO cB%
FOR k%=1 TO cA%
mat.c%(i%,j%) = mat.c%(i%,j%) + (mat.a%(i%,k%) * mat.b%(k%,j%))
NEXT k%
NEXT j%
NEXT i%
I also found a few more Basic-Plus2 MAT functions while
digging through the on-line help trying to confirm the
above. They are...
DET - Returns determinant of a matrix
NUM - Returns number of rows entered into a matrix with MAT INPUT
NUM2 - Returns number of elements in last row entered
TRN - Creates the transpose of a matrix
ptl
--- msged 2.07
* Origin: PTL Pointwork (1:105/48.111)
