Social Media Mining: An Introduction

(Axel Boer) #1

P1: qVa Trim: 6.125in×9.25in Top: 0.5in Gutter: 0.75in
CUUS2079-07 CUUS2079-Zafarani 978 1 107 01885 3 January 13, 2014 17:17


7.6 Bibliographic Notes 211

some benefits to themselves or is rational, and based on that, they will align
with the population. In herding, some level of uncertainty is associated with
the decision, and the individual does not know why he or she is following
the crowd.
Another confusion is that the terms “herd behavior/herding” is often used
interchangebly with “information cascades” [Bikhchandani et al., 1992;
Welch, 1992]. To avoid this problem, we clearly define both in the chapter
and assume that in herd behavior decisions are taken based on global
information, whereas in information cascades, local information is utilized.
Herd behavior has been studied in the context of financial markets [Cont
and Bouchaud, 2000;Drehmann et al., 2005;Bikhchandani and Sharma,
2001 ;Devenow and Welch, 1996] and investment [Scharfstein and Stein,
1990 ]. Gale analyzes the robustness of different herd models in terms of
different constraints and externalities [Gale, 1996], and Shiller discusses
the relation between information, conversation, and herd behavior [Shiller,
1995 ]. Another well-known social conformity experiment was conducted
in Manhattan byMilgram et al. [1969].
Other recent applications of threshold models can be found in [Young,
1988 ;Watts, 1999, 2002 ;Valente, 1995,1996a;Schelling, 1978;Peleg,
1997 ;Morris, 2000;Macy and Willer, 2002;Macy, 1991;Granovetter,
1976 ;Berger, 2001]. Bikhchandani et al. [1998] review conformity, fads,
and information cascades and describe how observing past human decisions
can help explain human behavior.Hirshleifer [1997] provides information
cascade examples in many fields, including zoology and finance.
In terms of diffusion models,Robertson [1967] describes the process and
Hagerstrand [1967] introduces a model based on the spatial stages of the
diffusion of innovations and Monte Carlo simulation models for diffusion of
innovations.Bass [1969] discusses a model based on differential equations.
Mahajan and Peterson [1978] extend the Bass model.
Instances of external-influence models can be found in [Hamblin et al.,
1973 ;Coleman et al., 1966] and internal-influence models are applied in
[Mansfield, 1961;Griliches, 2007;Gray, 2007]. The Gompertz function
[Martino, 1983], widely used in forecasting, has a direct relationship with
the internal-influence diffusion curve. Mixed-influence model examples
include the work ofMahajan and Muller [1982] and Bass model [Bass,
1969 ].
Midgley and Dowling [1978] introduce thecontingency model. Abra-
hamson et al.?mathematically analyze the bandwagon effect and diffusion Au:
“Abrahamson
1993” not found
in the reference
list. Please
provide.

of innovations. Their model predicts whether the bandwagon effect will
occur and how many organizations will adopt the innovation. Network
Free download pdf