The Classification Society
http://hdl.handle.net/2142/12230
Wed, 07 Oct 2015 12:37:58 GMT2015-10-07T12:37:58ZStatistical Issues in Agent Based Models
http://hdl.handle.net/2142/16518
Statistical Issues in Agent Based Models
Agent Based Models (ABMs) have become extremely popular tools for simulating certain kinds of complex phenomena. They have strong advantages in terms of both programming and validation. However, the statistical properties of the outputs of such models have not received significant theoretical attention. This talk will review the field, and then describe opportunities for research contributions.
Agent Based Models
Mon, 01 Jun 2009 00:00:00 GMThttp://hdl.handle.net/2142/165182009-06-01T00:00:00ZUltrametric Wavelet Regression of Multivariate Time Series: Application to Colombian Conflict Analysis
http://hdl.handle.net/2142/16517
Ultrametric Wavelet Regression of Multivariate Time Series: Application to Colombian Conflict Analysis
We first pursue the study of how hierarchy provides a well-adapted tool for the analysis of change. Then, using a time sequence-constrained hierarchical clustering, we develop the practical aspects of a new approach to wavelet regression. This provides a new way to link hierarchical relationships in a multivariate time series data set with external signals. Violence data from the Colombian conflict in the years 1990 to 2004 is used throughout. We conclude with some proposals for further study on the relationship between social violence and market forces, viz. between the Colombian conflict and the US narcotics market.
hierarchical clustering, wavelet regression, time series data
Mon, 01 Jun 2009 00:00:00 GMThttp://hdl.handle.net/2142/165172009-06-01T00:00:00ZSimultaneous Estimation of Multinomial Logistic Regression Models: Factor Analysis of Polytomous Item Response Data with Covariates
http://hdl.handle.net/2142/16516
Simultaneous Estimation of Multinomial Logistic Regression Models: Factor Analysis of Polytomous Item Response Data with Covariates
The approach described in this talk starts with Bock's (1972) nominal response model (NRM). The NRM is a multinomial logistic regression model for responses to items where the ordering of response options is not known a priori and the predictor or explanatory variable is unobserved (i.e., the latent construct). The latent variable in the multinomial logistic regression is replaced with an estimate based on responses to all other items. Given a set of items, there is one multinomial logistic regression for each. The problem then becomes one of simultaneously estimating multinomial logistic regressions with restrictions across the models. This approach allows us to go beyond what is typical in standard item response theory modeling in that we can handle multiple correlated latent constructs, impose (linear and/or ordinal) restrictions on category scores, test the effect of additional variables (e.g., a treatment versus control condition), create hybrid IRT models, and obtain measures on the latent constructs. An extension of the approach will also be described that is akin to a structural equation model for observed discrete data.
Multinomial Logistic Regression Factor Analysis
Mon, 01 Jun 2009 00:00:00 GMThttp://hdl.handle.net/2142/165162009-06-01T00:00:00ZDocument Clustering and Social Networks
http://hdl.handle.net/2142/12608
Document Clustering and Social Networks
Text Mining has become a specialized offshoot of Data Mining, Information Retrieval, and Natural Language Processing. One of the major tools of this area is the vector space representation of documents. On the other hand, social network analysis has found its mathematical underpinnings primarily in mathematical graph theory. A graph has a dual representation as an adjacency matrix. So-called two-mode social networks have actors of two different types, frequently individuals and organizations. The adjacency matrix for these two-mode social networks has the same structure as the so-called term-document matrices used in text mining. In the talk we discuss these connections and show how these ideas can be exploited in both fields. In particular, methods for block modeling in social network analysis can be used for document clustering.
Cluster analysis, text mining, mixture models, social networks
Mon, 01 Jun 2009 00:00:00 GMThttp://hdl.handle.net/2142/126082009-06-01T00:00:00Z