A quadratic lower bound for Rocchio’s similarity-based relevance feedback algorithm with a fixed query updating factor

Rocchio’s similarity-based relevance feedback algorithm, one of the most important query reformation methods in information retrieval, is essentially an adaptive supervised learning algorithm from examples. In practice, Rocchio’s algorithm often uses a fixed query updating factor. When this is the case, we strengthen the linear Ω(n) lower bound obtained by Chen and Zhu (Inf. Retr. 5:61–86, 2002) and prove that Rocchio’s algorithm makes Ω(k(n−k)) mistakes in searching for a collection of documents represented by a monotone disjunction of k relevant features over the n-dimensional binary vector space {0,1} ⁿ , when the inner product similarity measure is used. A quadratic lower bound is obtained when k is linearly proportional to n. We also prove an O(k(n−k)³) upper bound for Rocchio’s algorithm with the inner product similarity measure in searching for such a collection of documents with a constant query updating factor and a zero classification threshold.     

Poverty Assessment in Degraded Rural Drylands in the Monte Desert,Argentina. An Evaluation Using GIS and Multi-criteria Decision Analysis

The work aims to contribute to the construction a local-scale poverty indicator, which contemplates multiple dimensions and allows for spatialization of socioeconomic data for a rural area in the Monte Desert. Given the evident and widely studied relationship between desertification processes and poverty, and aiming to contribute to integrating socioeconomic information to desertification assessment and monitoring, there arises the need for spatializing poverty by addressing its multiple dimensions, aspects poorly developed thus far. For this purpose, a data model was designed, which enabled integrating quantitative and qualitative information within the scope of geographic ?nformation systems (GIS) with multi-criteria decision analysis (MCDA), whereby it was possible to spatialize in detail the degree of poverty in the study area, laying the foundations for up-scaling the assessment to different scales. This work demonstrates the usefulness of GIS and MCDA as an instrument that enables progressing in new integral, interdisciplinary, multi-scale and multi-temporal approaches.     

Why reputation is not always beneficial: Tolerance and opportunism in business networks

Many researchers in economics as well as sociology have stressed the important role of business networks for cooperation, trust, and performance. This claim is based on solid theoretical arguments as well as empirical findings. However, neither theory nor the selective empirical results support the view that networks are always beneficial for economic transactions. This paper begins with an observation that for the purchase of IT products, network embeddedness leads to even more problems for the customer. In order to explain this effect, possible reasons for this phenomenon are discussed using theory as well as empirics. The most promising explanation for this special case is the effect of uncertainty and incomplete information ex post. In order to reduce this uncertainty, the buyer forms beliefs on the basis of the opinions existing in a shared network or group. However, if the network members have the same problem of uncertainty, suppliers have an incentive to reduce their performance because such behavior will not be detected and sanctioned. An analysis of the customer's tolerance to a supplier's behavior in business transactions yields support for this argument. Even if problems in a transaction are kept constant, customers give suppliers more reputational credit if they share a common network.     

Migration Decisions Within Dual‐Earner Partnerships: A Test of Bargaining Theory

This article focuses on the problems faced by dual‐earner partnerships arising from the regional coordination of the partners' careers. It is still unknown whether the fact that couples are less mobile than singles is caused by homogeneous preferences within couples or by a process of balancing conflicting interests. Consequently, we analyze the potential conflicts provoked by work‐related migration incentives. We test hypotheses derived from bargaining theory using quasi‐experimental data from a factorial survey of nearly 280 European couples. Our results support the bargaining approach and confirm that asymmetrical shifts in bargaining power drive the potential for conflict. Women's willingness to move is generally less than men's, but the impact of employment prospects differs only slightly by gender.     

Modelling state anxiety in a medical setting: The mediating role of embarrassment

Abstract

Little research has been conducted on the relationship between state anxiety and embarrassment in a medical setting. In the prcsent study, several personality and situational variables were examined as antecedents of state anxiety. A total of 277 patients who had come for an examination either as a response to a melanoma screening programme or due to a dermatological complaint participated in the study. Results showed that purpose of patient visit, physician's recommendation, trait anxiety, and embarrassment, were each correlated with state anxiety. Furthermore, using mediated regression analysis as well as structural equations, embarrassment was found to mediate between the above antecedents and state anxiety. The authors proposed a model linking all the variables and identifying their paths. Theoretical and practical considerations of the findings are discussed.     

On characterizing mixtures of some life distributions

Received: October 15, 1998; revised version: January 10, 2000     

A test of the Core,Bargaining Set,Kernel and Shapley models in n-person quota games with one weak player

n-person (n – 1)-quota-games where the quotas are positive for certain n – 1 (strong) players and negative for the remaining (weak) player, are discussed. Normative solutions predicted by the Core,the Kernel, the Bargaining Set, the Competitive Bargaining Set, and by the Shapley Value are presented and exemplified.Each of twelve groups of subjects participated in a four-person and a five-person (n – 1)-quota games with one weak player. The weak player was always excluded from the ratified coalition. The division of payoffs among the strong players was more egalitarian than the Kernel solution but less egalitarian than the Shapley value. The Core and the Bargaining Sets were fully supported for the two strongest players, but less supported for the other players. Analyses of the bargaining process confirmed a dynamic interpretation of the Bargaining Set Theory.This research was performed while the author was at the University of North Carolina. The research was partially supported by a PHS Research Grant No. MH-10006 from the National Institute of Mental Health. The author thanks Professor Amnon Rapoport for helpful advice in the design of this study.     

Empirical likelihood based variable selection

Information criteria form an important class of model/variable selection methods in statistical analysis. Parametric likelihood is a crucial part of these methods. In some applications such as the generalized linear models, the models are only specified by a set of estimating functions. To overcome the non-availability of well defined likelihood function, the information criteria under empirical likelihood are introduced. Under this setup, we successfully solve the existence problem of the profile empirical likelihood due to the over constraint in variable selection problems. The asymptotic properties of the new method are investigated. The new method is shown to be consistent at selecting the variables under mild conditions. Simulation studies find that the proposed method has comparable performance to the parametric information criteria when a suitable parametric model is available, and is superior when the parametric model assumption is violated. A real data set is also used to illustrate the usefulness of the new method.     

Estimation in Epidemics with Incomplete Observations

The construction of estimating equations by martingale methods is generalized to yield estimators with explicit expressions for the parameters of the birth-and-death and the general epidemic processes when only partial observations are available. (For the birth-and-death process the death process is observed but the number of births is observed only at the end and for the general epidemic process only the removal process is observed.) For large populations, the use of the martingale central limit theorem yields asymptotic confidence regions for the parameters. Explicit expressions are derived for estimators of the variances of the large sample distributions. The range of validity and usefulness of the new estimators is determined by simulation.     

Forecast efficiency of systematically sampled time series

In this paper we consider the effect of the sampling interval in systematically sampled time series on the forecast efficiency. We find that for sufficientlv long sampled series the loss in forecasting accuracy by using the sampled instead of the original series is small