8:26 p.m.
I'm responding to several recent posts by Sven and Ioannis.
1) I accept Sven's point that "eiggen2" is not a great name for the
revised (to deal with native complex matrices) variant of the old
"eigengen" function. In git, I've now renamed eiggen2() to eigen(),
and revised the doc accordingly. But for the present I've retained
"eiggen2" as an alias since I was concerned about breaking the
ghosts and/or DFM packages.
2) I accept Ioannis's point that the new functions Re() and Im()
might reasonably be applied to real matrices as well as complex
ones. Given real input, Re() would return a copy of the input and
Im() would return an appropriately sized zero matrix. That too is
now in git.
3) I'm not so sure about Sven's suggestion that (if I'm
understanding him right) sqrt(M), for M a real matrix with some
negative elements, should return an appropriate complex matrix. It's
clearly defensible, but it might produce surprising/puzzling results
in some cases so I think it requires further discussion.
Allin
9:02 p.m.
On Fri, 10 Jan 2020, Allin Cottrell wrote:
3) I'm not so sure about Sven's suggestion that (if I'm
understanding him
right) sqrt(M), for M a real matrix with some negative elements, should
return an appropriate complex matrix. It's clearly defensible, but it might
produce surprising/puzzling results in some cases so I think it requires
further discussion.
In the meantime, it's not difficult to get what you want in this
sort of case, if you know what you're doing. Example:
<hansl>
matrix B = {1,-4; -9,1}
print B
eval sqrt(complex(B,0))
</hansl>
<output>
B (2 x 2)
1 -4
-9 1
? eval sqrt(complex(B,0))
1 + 0i 0 + 2i
0 + 3i 1 + 0i
</output>
Allin
9:08 p.m.
Am 10.01.2020 um 22:02 schrieb Allin Cottrell:
On Fri, 10 Jan 2020, Allin Cottrell wrote:
> 3) I'm not so sure about Sven's suggestion that (if I'm understanding
> him right) sqrt(M), for M a real matrix with some negative elements,
> should return an appropriate complex matrix. It's clearly defensible,
> but it might produce surprising/puzzling results in some cases so I
> think it requires further discussion.
First of all, yes I understand what you mean, it would have some
downsides, too.
In the meantime, it's not difficult to get what you want in this
sort
of case, if you know what you're doing. Example:
<hansl>
matrix B = {1,-4; -9,1}
print B
eval sqrt(complex(B,0))
</hansl>
I was thinking of the same approach, but didn't know that a simple 0 was
enough and would broadcast. In the doc it suggests instead something
like zeros(2,2). To be honest, I would still make the second arg optional...
thanks for the other changes,
sven
11:29 p.m.
On Fri, 10 Jan 2020, Sven Schreiber wrote:
Am 10.01.2020 um 22:02 schrieb Allin Cottrell:
> On Fri, 10 Jan 2020, Allin Cottrell wrote:
>
>> 3) I'm not so sure about Sven's suggestion that (if I'm
understanding
>> him right) sqrt(M), for M a real matrix with some negative elements,
>> should return an appropriate complex matrix. It's clearly defensible,
>> but it might produce surprising/puzzling results in some cases so I
>> think it requires further discussion.
>
First of all, yes I understand what you mean, it would have some
downsides, too.
> In the meantime, it's not difficult to get what you want in this sort
> of case, if you know what you're doing. Example:
>
> <hansl>
> matrix B = {1,-4; -9,1}
> print B
> eval sqrt(complex(B,0))
> </hansl>
I was thinking of the same approach, but didn't know that a simple 0 was
enough and would broadcast.
That's actually quite explicit in the User's Guide, section 17.2.
But admittedly the online help for complex() is not so forthcoming;
amendment needed.
Allin
9:03 p.m.
On Fri, 10 Jan 2020, Allin Cottrell wrote:
3) I'm not so sure about Sven's suggestion that (if I'm
understanding him
right) sqrt(M), for M a real matrix with some negative elements, should
return an appropriate complex matrix. It's clearly defensible, but it might
produce surprising/puzzling results in some cases so I think it requires
further discussion.
This is what R does:
<R>
sqrt(as.matrix(c(1, -2)))
[,1]
[1,] 1
[2,] NaN
</R>
and this is what Octave (therefore, I presume, Matlab) does:
<octave>
> sqrt([1 -2])
ans =
1.00000 + 0.00000i 0.00000 + 1.41421i
</octave>
Personally, I'd stick with the R approach.
-------------------------------------------------------
Riccardo (Jack) Lucchetti
Dipartimento di Scienze Economiche e Sociali (DiSES)
Università Politecnica delle Marche
(formerly known as Università di Ancona)
r.lucchetti(a)univpm.it
http://www2.econ.univpm.it/servizi/hpp/lucchetti
-------------------------------------------------------
11:24 p.m.
On Fri, 10 Jan 2020, Riccardo (Jack) Lucchetti wrote:
On Fri, 10 Jan 2020, Allin Cottrell wrote:
> 3) I'm not so sure about Sven's suggestion that (if I'm understanding him
> right) sqrt(M), for M a real matrix with some negative elements, should
> return an appropriate complex matrix. It's clearly defensible, but it might
> produce surprising/puzzling results in some cases so I think it requires
> further discussion.
This is what R does:
<R>
> sqrt(as.matrix(c(1, -2)))
[,1]
[1,] 1
[2,] NaN
</R>
and this is what Octave (therefore, I presume, Matlab) does:
<octave>
>> sqrt([1 -2])
ans =
1.00000 + 0.00000i 0.00000 + 1.41421i
</octave>
Personally, I'd stick with the R approach.
Me too.
Allin
1808
days inactive
1808
days old
5 comments
3 participants
participants (3)
-
Allin Cottrell -
Riccardo (Jack) Lucchetti -
Sven Schreiber