This is website http://www.stats.ox.ac.uk/~snijders/stat_net_mod.htm.
This website contains the announcements of an informal group meeting irregularly in Nuffield College about the topic of statistical models for social networks. Anybody interested is welcome.
Organized by Tom Snijders and Johan Koskinen.
Meetings in TT 2014:
Roberto Franzosi, a visitor of the Sociology Group in Nuffield, talked about Quantitative Narrative Analysis (QNA), which can be used also for analyzing networks. His interest is in combining a network approach with Geographical Information Systems (GIS).
Abstract
Quantitative Narrative Analysis (QNA) is a computer-assisted approach
to narrative based on invariant linguistic properties of narrative.
Basically, a narrative clause is structured around the 5 Ws of journalism + H:
who, what, when, where, why and how (a jingle already known in classical
rhetoric as quis, quid, quando, ubi, cur, quem ad modum, quibus adminiculis).
It is the who-what-who combination that makes QNA data amenable
to network graphs (dynamic graphs with the addition of the when)
while the who-what-where makes possible the use of GIS models (again,
dynamic with the addition of when) with data that are basically words.
I will illustrate these data visualization problems using QNA data
gathered from thousands of newspaper articles on the rise of Italian
fascism (1919-1922) and on lynchings of African-Americans in Georgia (1875-1930).
Meetings in MT 2013:
Meetings in HT 2010:
Meetings held in TT 2009:
Meetings in TT 2013:
Discussion about whatever new developments have come up.
Presentations:
Speaker: Robert Hellpap.
No meeting!
James Hollway: presentation/discussion.
Presenters: Per Block; Cohen Simson.
There will be meetings in weeks 4 and 8.
Meetings in MT 2012:
Presentations by Per Block, Zsofia Boda, Andras Voros, and James Hollway
(preview and discussion of Sunbelt presentations).
Discussion about progress of RSiena: content, coding, documentation.
Meetings in TT 2012:
First meeting after the summer break.
This will be the last meeting, for some time to come, that Josh Lospinoso is with us.
We also shall have a guest from Konstanz, Mark Ortmann.
The meeting will be devoted to exchanging our network modeling progress over the summer,
and discussing with Mark about his plans.
Tom Snijders: using
sienaGOF
and perhaps other RSiena goodies.
Meetings in HT 2012:
Per Block: Gender-specific influence among adolescents with respect to
smoking and drinking behaviour.
3.45-5.00 pm. RSiena coding session.
Gareth Cochrane (Belfast) on the Belfast Youth Development Study.
3.45-5.00 pm. RSiena coding session.
Charlotte Greenan on estimation algorithms (as alternative to Robbins-Monro).
3.45-5.00 pm. RSiena coding session.
Nynke Niezink: Models for co-evolution of networks and behavior for
continuous behavior variables, modeled by stochastic differential equations.
3.45-5.00 pm. RSiena coding session.
Meetings in MT 2011:
Charlotte Greenan: Innovation diffusion in social networks.
3.45-5.00 pm. RSiena coding session: new effects in RSiena, unit testing.
András Vörös.
3.45-5.00 pm. RSiena coding session.
Who will speak is still open - Tom, if nobody else wants to.
3.45-5.00 pm. RSiena coding session.
2.00-2.45 pm. Free discussion.
3.45-5.00 pm. RSiena coding session.
Meetings in TT 2011:
Johan Koskinen, co-author Sten-Åke Stenberg,
Models for multilevel analysis of binary response with peer
dependencies.
Per Block: Multiple identities in the evolution of small social networks.
Griffith Rees: Interval censoring with Cox models for spatial diffusion.
3.45-5.00 pm. RSiena code session.
Discussion about the practical use of the proposals by Natalie Indlekofer
(from her talk at the Nuffield-OII social networks seminar on October 24)
for parameter interpretation: Which proposals should mainly be recommended
for use in applications of Siena?
3.45-5.00 pm. RSiena code session.
2.00-2.45 pm. Paulina Preciado (preview of talk to be given later in Manchester).
2.45-3.30 pm. Natalie Indlekofer (report about work she has been doing in Oxford about
expressing influence of individual actors on estimates in actor-oriented models).
3.45-5.00 pm. RSiena code session.
Meetings in HT 2011:
(Note: Tom speaks at the Nuffield Sociology Seminar, Wednesday May 11, 5pm, about
Micro mechanisms and macro structures of social networks.)
Amber Thomas.
Tom Snijders about models for multiple dependent networks.
Steve Manion about his network study of Wikipedia.
Meetings in MT 2010:
Charlotte Greenan.
Meetings in TT 2010:
Ainhoa de
Federico (University of Toulouse - II Le Mirail) and Clair Bidart
(CNRS, Laboratoire d'Economie et de Sociologie du Travail,
Aix-en-Provence, France):
Analysis of personal networks I
Miranda Lubbers
Analysis of personal networks II
Josh Lospinoso. Analysis of relational events and time-stamped network data.
(Names of presenters are not unchangeable; topics to be added.)
Josh Lospinoso.
Johan Koskinen.
Ruth Ripley.
Meetings held in MT 2009:
Tom Snijders: Models for dynamics of non-directed networks.
This will be mainly about Section 3 of
the paper Actor-Based Models for Network Dynamics
(recently submitted).
Paulina Preciado Lopez: Distance matters:
exploratory analysis of the distance dependency of adolescent friendship dynamics.
Krista Gile: Inference for Hidden Populations Based on Network Sampling
Josh Lospinoso: Assessing and correcting parametrizations for time heterogeneity
in stochastic actor-based models.
Peng Wang (University of Melbourne): the design and source code
of
PNet.
Philippa Pattison (University of Melbourne):
Some thoughts on specifying and
estimating exponential random graph models for large networks.
Issues related to Network-Based Respondent-Driven Sampling.
Part 1: Amber Tomas: sensitivity of estimators to network and sampling features;
Part 2: Krista Gile: a new estimator based on a network model.
Statistical Network Modeling :
this was a full-day workshop
starting 9.00.
Nuffield College, Seminar Room.
The programme can be seen here.
Viviana Amati: Estimation
in Actor-Oriented Network Models by Simulated Generalized Method of Moments.
Johan Koskinen: Extreme actors in ERGMs - towards a model-based centrality measure?
This talk is based on tech report
MelNet_Techreport_08_05.pdf.
The meeting will start with the discussion about Johan's paper (see above),
for which we had no time on May 4.
Next in this meeting, Bernie Hogan will talk about
Real-time network layouts, community detection and alter sampling
in personal interviews: Strategies and techniques using the Facebook API.
A related working paper is
A Comparison
of On and Offline Networks through the Facebook API.
This talk and discussion will be about data collection methods
rather than statistical modeling issue, but
the organizers like to have a broad definition of this group.
There will be three topics: