This is a multi-part message in MIME format.
------=_NextPart_000_0001_01C31AB9.44A9B370 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: quoted-printable Hi all, I've attached the notes from Mohammed's & my online chat re validation & methodology. The main points discussed were: 1. Distinguish between (i) academics' and (ii) consultants' use of validation & methodology. Former through peer review, being seen to do = the work 'the right way'; latter through whether it gives the right results = or not. The overriding aims of (i) is to try to have a "basic" = understanding about such varied topics as the economy, modes of social interactions, population dynamics, computational anthropology, etc.. while on (ii) is = has more to do with trying to make better decisions that will have a direct effect on the bottom over some limited time horizon.=20 2. Misuse of ABMs:=20 * the misinterpratetion of many-to-one problems (in particular, when attempting to induce (micro) rules from observed (macro) behaviour;=20 * Over claiming predictive abilities (clients expecting quantitiative predictions and getting qualitative ones).=20 * Very limited attempts at validation with the real world.=20 3. Proper use of ABMs: at least one proper use (both our experience) is starting from (micro) rules and assumptions and generating the (macro) behaviour that they imply. This deductive process is one-to-one in = contrast to the inductive many-to-one. A great deal of value of ABM to the = business world is in helping them understand more about the fallacy/veracity of = their assumptions. This in itself would help them make better decisions = without claims of predictive accuracy.=20 4. How to proceed:=20 * Ultimately, funding is needed=20 * Involvement of credible parties (SFI/BizNet?) (probably not the established consultancies - either (i) already have their own approach = or (ii) would be embarassed to admit to clients that they didn't.=20 * Assemble a 'version 0.1' framrework - a brain-dump of as many practitioners' approaches as practical.=20 5. Actions:=20 * RH to explore SFI/BizNet aspect * Both to carry out their brain-dumps, try and get others to do same and attempt to incorporate. (If anyone has any ideas about how to do = this, please pitch in). Robert ------=_NextPart_000_0001_01C31AB9.44A9B370 Content-Type: text/html; charset="us-ascii" Content-Transfer-Encoding: quoted-printable <!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 3.2//EN"> <HTML> <HEAD> <META HTTP-EQUIV=3D"Content-Type" CONTENT=3D"text/html; = charset=3Dus-ascii"> <META NAME=3D"Generator" CONTENT=3D"MS Exchange Server version = 6.0.4630.0"> <TITLE>V&M notes</TITLE> </HEAD> <BODY> <!-- Converted from text/rtf format --> <P><FONT SIZE=3D2 FACE=3D"Trebuchet MS">Hi all,</FONT> </P> <P><FONT SIZE=3D2 FACE=3D"Trebuchet MS">I've attached the notes from = Mohammed's & my online chat re validation & methodology. The = main points discussed were:</FONT> <BR><SPAN LANG=3D"en-gb"><FONT SIZE=3D2 FACE=3D"Trebuchet MS">1. = Distinguish between (i) academics' and (ii) consultants' use of = validation & methodology. Former through peer review, being seen to = do the work 'the right way'; latter through whether it gives the right = results or not. The overriding aims of (i) is to try to have a = "basic" understanding about such varied topics as the economy, = modes of social interactions, population dynamics, computational = anthropology, etc.. while on (ii) is has more to do with trying to make = better decisions that will have a direct effect on the bottom over some = limited time horizon. </FONT></SPAN></P> <P><SPAN LANG=3D"en-gb"><FONT SIZE=3D2 FACE=3D"Trebuchet MS">2. Misuse = of ABMs: </FONT></SPAN> <UL> <UL> <LI><SPAN LANG=3D"en-gb"><FONT SIZE=3D2 FACE=3D"Trebuchet MS">the = misinterpratetion of many-to-one problems (in particular, when = attempting to induce (micro) rules from observed (macro) behaviour; = </FONT></SPAN></LI> <LI><SPAN LANG=3D"en-gb"><FONT SIZE=3D2 FACE=3D"Trebuchet MS">Over = claiming predictive abilities (clients expecting quantitiative = predictions and getting qualitative ones). </FONT></SPAN></LI> <LI><SPAN LANG=3D"en-gb"><FONT SIZE=3D2 FACE=3D"Trebuchet MS">Very = limited attempts at validation with the real world. </FONT></SPAN></LI> </UL></UL> <P><SPAN LANG=3D"en-gb"><FONT SIZE=3D2 FACE=3D"Trebuchet MS">3. Proper = use of ABMs: at least one proper use (both our experience) is starting = from (micro) rules and assumptions and generating the (macro) behaviour = that they imply. This deductive process is one-to-one in contrast to the = inductive many-to-one. A great deal of value of ABM to the business = world is in helping them understand more about the fallacy/veracity of = their assumptions. This in itself would help them make better decisions = without claims of predictive accuracy. </FONT></SPAN></P> <P><SPAN LANG=3D"en-gb"><FONT SIZE=3D2 FACE=3D"Trebuchet MS">4. How to = proceed: </FONT></SPAN> <UL> <UL> <LI><SPAN LANG=3D"en-gb"><FONT SIZE=3D2 FACE=3D"Trebuchet = MS">Ultimately, funding is needed </FONT></SPAN></LI> <LI><SPAN LANG=3D"en-gb"><FONT SIZE=3D2 FACE=3D"Trebuchet = MS">Involvement of credible parties (SFI/BizNet?) (probably not the = established consultancies - either (i) already have their own approach = or (ii) would be embarassed to admit to clients that they didn't. = </FONT></SPAN></LI> <LI><SPAN LANG=3D"en-gb"><FONT SIZE=3D2 FACE=3D"Trebuchet MS">Assemble a = 'version 0.1' framrework - a brain-dump of as many practitioners' = approaches as practical. </FONT></SPAN></LI> </UL></UL> <P><SPAN LANG=3D"en-gb"><FONT SIZE=3D2 FACE=3D"Trebuchet MS">5. Actions: = </FONT></SPAN> <UL> <LI><SPAN LANG=3D"en-gb"><FONT SIZE=3D2 FACE=3D"Trebuchet MS">RH to = explore SFI/BizNet aspect</FONT></SPAN></LI> <LI><SPAN LANG=3D"en-gb"><FONT SIZE=3D2 FACE=3D"Trebuchet MS">Both to = carry out their brain-dumps, try and get others to do same and attempt = to incorporate. (If anyone has any ideas about how to do this, please = pitch in).</FONT></SPAN></LI> <BR> </UL> <P><SPAN LANG=3D"en-gb"><FONT SIZE=3D2 FACE=3D"Trebuchet = MS">Robert</FONT></SPAN> </P> </BODY> </HTML> ------=_NextPart_000_0001_01C31AB9.44A9B370-- |
Free forum by Nabble | Edit this page |