From owner-chemistry@ccl.net Tue Jan 12 14:21:01 2021 From: "wdlaidig()gmail.com" To: CCL Subject: CCL: Least Trimmed Squares versus All Possible Subsets Message-Id: <-54257-210112101318-18088-aJeq1lbrZLMLtbyzuefOHA!=!server.ccl.net> X-Original-From: Content-Language: en-us Content-Type: multipart/alternative; boundary="----=_NextPart_000_05EA_01D6E8CB.89458C90" Date: Tue, 12 Jan 2021 10:13:12 -0500 MIME-Version: 1.0 Sent to CCL by: [wdlaidig . gmail.com] This is a multipart message in MIME format. ------=_NextPart_000_05EA_01D6E8CB.89458C90 Content-Type: text/plain; charset="utf-8" Content-Transfer-Encoding: quoted-printable Hi John, =20 I am not sure what APS stands for, but assuming it stands for all = subsets possible don=E2=80=99t know if I would call it a subset of least = trimmed squares. I am thinking that at each k <=3D n length, APS finds = the minimizes objective function value for LTS values for that k. = However, LTS, I believe, uses all n points to evaluate the final least = squares value instead of just the k best point set. You should be able = to get the LST value by using the equation determined for subset length = k in APS and calculating least squares using all n points (analogous to = how LS is calculated from the least median square line). =20 Take care, Bill =20 > From: owner-chemistry+wdlaidig=3D=3Dgmail.com[a]ccl.net = On Behalf Of John = McKelvey jmmckel_._gmail.com Sent: Monday, January 11, 2021 10:39 AM To: Laidig, Bill Subject: CCL: Least Trimmed Squares versus All Possible Subsets =20 Folks, =20 Would it be reasonable to consider APS, as found in MINITAB, as a subset = of LTS? =20 I would appreciate knowing of a user forum in the area of Statistics. =20 Many thanks in advance, =20 John McKelvey =20 =20 =20 =20 =20 --=20 John McKelvey 545 Legacy Pointe Dr O'Fallon, MO 63376 636-294-5203 jmmckel_._gmail.com =20 ------=_NextPart_000_05EA_01D6E8CB.89458C90 Content-Type: text/html; charset="utf-8" Content-Transfer-Encoding: quoted-printable

Hi John,

 

I am not = sure what APS stands for, but assuming it stands for all subsets = possible don=E2=80=99t know if I would call it a subset of least trimmed = squares.=C2=A0 I am thinking that at each k <=3D n length, APS finds = the minimizes objective function value for LTS values for that k.=C2=A0 = However, LTS, I believe, uses all n points to evaluate the final least = squares value instead of just the k best point set.=C2=A0 You should be = able to get the LST value by using the equation determined for subset = length k in APS and calculating least squares using all n points = (analogous to how LS is calculated from the least median square = line).

 

Take care,

Bill

 

From: = owner-chemistry+wdlaidig=3D=3Dgmail.com[a]ccl.net = <owner-chemistry+wdlaidig=3D=3Dgmail.com[a]ccl.net> On Behalf Of = John McKelvey jmmckel_._gmail.com
Sent: Monday, January = 11, 2021 10:39 AM
To: Laidig, Bill = <wdlaidig[a]gmail.com>
Subject: CCL: Least Trimmed Squares = versus All Possible Subsets

 

Folks,

 

Would it be reasonable to consider APS, as found in = MINITAB, as a subset of LTS?

 

I = would appreciate knowing of a user forum in the area of =  Statistics.

 

Many thanks in advance,

 

John McKelvey

 

 

 

 

 

-- =

John = McKelvey
545 Legacy Pointe Dr

O'Fallon, MO 63376
636-294-5203
jmmckel_._gmail.com

------=_NextPart_000_05EA_01D6E8CB.89458C90--