Interpret input quantity matrix multiplied by the transpose of a price matrix

68 Views Asked by At

I have two $77 \times 8$ matrices where rows represent firm ids:

  • input quantity matrix, $X$, where columns represent quantity types (e.g., grain, seed, and chemical).

  • price matrix, $W$, where columns represent prices of the corresponding quantities from $X$.

I want to calculate the total cost for each id but since the dimensions of the matrices are non-conformable I take the transpose of $W$ and calculate $$X W^{T}$$ which is a $77 \times 77$ matrix. How do I interpret the result? If I've done this incorrectly then how do I identify the total cost for each id?


Example

The matrix $X$:

       Quantity 1 Quantity 2 Quantity 3
Firm 1  13.727112  11.134628   6.898464
Firm 2   3.252877  17.258291   7.429559
Firm 3   8.793248   5.896631   9.765461
Firm 4  11.943213  13.869619  14.947700
Firm 5  17.133765  16.214109   6.019190
Firm 6  16.390385  10.711969  12.892855
Firm 7  12.650650   4.110215  14.750364

The matrix $W$:

        Price 1   Price 2  Price 3
Firm 1 1.684008 2.9795928 2.664915
Firm 2 2.423838 2.6259991 2.289327
Firm 3 2.579981 1.9127199 2.229167
Firm 4 2.450678 2.9854698 1.705296
Firm 5 2.099541 0.8551909 2.632429
Firm 6 1.375625 1.7132088 2.287841
Firm 7 2.429695 1.2428014 2.731882
1

There are 1 best solutions below

0
On

Using the Hadamard product,

$$\mbox{diag}^{-1} \left( {\bf X} {\bf W}^\top \right) = \left( {\bf X} \circ {\bf W} \right) {\bf 1}_8$$

as mentioned in equation $(73)$ of Minka's Old and new Matrix Algebra useful for Statistics.