[Date Prev][Date Next][Thread Prev][Thread Next][Date index][Thread index]

From |
"Andre pierre" <hcpats@hotmail.fr> |

To |
statalist@hsphsun2.harvard.edu |

Subject |
RE: st: RE: grouping by time window |

Date |
Mon, 13 Jun 2005 16:17:09 +0200 |

Thanks for your comments. There is indeed an argument I didn't mention to justify this way of treating the data (that is: inventions produced at very closed dates, by the same firm are identical, or at least very similar in my data). I thus have to choose an arbitrary time window (that could be modified) to identify these "portfolios".

I should ideally find a measure that maximises the number of products in each portfolio in a given time window, but that seems rather difficult to implement...

André

_________________________________________________________________From: "Nick Cox" <n.j.cox@durham.ac.uk> Reply-To: statalist@hsphsun2.harvard.edu To: <statalist@hsphsun2.harvard.edu> Subject: st: RE: grouping by time window Date: Mon, 13 Jun 2005 14:35:13 +0100 I am not sure this is determinate on the criteria given. In particular, if you start a first window at the first observation and then use the rule that no window can be more than a year long, then the windows so produced do not necessarily coincide with the set of disjoint windows that maximise the number of products within each, if indeed there is a unique set so defined. A quite different comment -- and here you may well have a rationale not presented -- is that seems a rather arbitrary way to reduce these data. Nick n.j.cox@durham.ac.uk Andre pierre > I would like to create groups of observations with at least 2 > products > invented within 365 days (not a civil year), by firm and > industry code. > > obs; Firm; ic; date > 1; 1; A; 31 dec 02 > 2; 1; A; 2 Apr 03 > 3; 1; A; 12 dec 03 > 4; 1; A; 5 jan 04 > 5; 1; A; 4 may 04 > 6; 1; B; 4 may 04 > 7; 2; A; 1 jan 01 > . > . > . > > > In this example, observations 1 to 3 and 4 to 5 would correspond to a > "group", as the products are invented by the same firm, in > the same industry > code within 365 days. The groups have to be "successive", and not > "overlapping". * * For searches and help try: * http://www.stata.com/support/faqs/res/findit.html * http://www.stata.com/support/statalist/faq * http://www.ats.ucla.edu/stat/stata/

MSN Search : des réponses à tous vos besoins ! http://www.imagine-msn.com/hotmail/default.aspx?locale=fr-FR

*

* For searches and help try:

* http://www.stata.com/support/faqs/res/findit.html

* http://www.stata.com/support/statalist/faq

* http://www.ats.ucla.edu/stat/stata/

**References**:**st: RE: grouping by time window***From:*"Nick Cox" <n.j.cox@durham.ac.uk>

- Prev by Date:
**st: RE: grouping by time window** - Next by Date:
**Re: st: apple to intel** - Previous by thread:
**st: RE: grouping by time window** - Index(es):

© Copyright 1996–2017 StataCorp LLC | Terms of use | Privacy | Contact us | What's new | Site index |