作者: Daniel A. Griffith
DOI: 10.1007/978-3-642-01554-0_16
关键词:
摘要: Carey (1858) and Ravenstein (1885) first proposed, through analogy, the gravity model of Newtonian physics as a description for economic social spatial interaction, with Sen Smith (1995) furnishing comprehensive treatment this more than century later. In late 1960s, Wilson spelled out an entropy maximizing derivation model, including use row column totals additional information modeling purposes (that is, doubly-constrained version), followed by utility maximization it Niedercorn Bechdolt (1969). Flowerdew Atkin (1982) Lovett (1988) articulated linkages between Poisson probability interaction. Within same time interval, Anas (1983) established linkage logit joint origin-destination choice, which indirectly relates to specification that includes separate indicator variable each origin destination 2n 0–1 binary variables, having single 1 n-1 0s). Curry (1972; also see et al. 1975, 1976) Griffith Jones (1980), raised issue autocorrelation effects embedded in These investigations were formulation network concept (see Black 1992; Thomas 1998; Tiefelsdorf Braun 1999). More recently, LeSage Pace (2008), Fischer (2008) have returned specifying autoregressive filter versions unconstrained but terms attribute geographic distributions. Chun (2007) moves beyond conceptualization explicitly spatially autocorrelated flows.