Optimal ensemble construction via meta-evolutionary ensembles

In this paper, we propose a meta-evolutionary approach to improve on the performance of individual classifiers. In the proposed system, individual classifiers evolve, competing to correctly classify test points, and are given extra rewards for getting difficult points right. Ensembles consisting of multiple classifiers also compete for member classifiers, and are rewarded based on their predictive performance. In this way we aim to build small-sized optimal ensembles rather than form large-sized ensembles of individually-optimized classifiers. Experimental results on 15 data sets suggest that our algorithms can generate ensembles that are more effective than single classifiers and traditional ensemble methods.     

Mapping the semantics of Web text and links

Search engines use content and links to search, rank, cluster, and classify Web pages. These information discovery applications use similarity measures derived from this data to estimate relatedness between pages. However, little research exists on the relationships between similarity measures or between such measures and semantic similarity. The author analyzes and visualizes similarity relationships in massive Web data sets to identify how to integrate content and link analysis for approximating relevance. He uses human-generated metadata from Web directories to estimate semantic similarity and semantic maps to visualize relationships between content and link cues and what these cues suggest about page meaning. Highly heterogeneous topical maps point to a critical dependence on search context.     

Cancer incidence in a cohort of infertile women

Among 2,496 infertile Israeli women treated between 1964 and 1974, 143 cancer cases were observed as compared with 116.1 expected (standardized incidence ratio (SIR) = 1.2, 95% confidence interval (CI) 1.0-1.5) through 1991. Site-specific analysis revealed 12 ovarian cancers versus 7.2 expected (SIR = 1.6, 95% CI 0.8-2.9), 21 endometrial cancers versus 4.3 expected (SIR = 4.85, 95% CI 3.0-7.4), and 59 breast cancers versus 46.6 expected (SIR = 1.3, 95% CI 0.96-1.6). Sensitivity analysis revealed that confounding was unlikely to explain the raised risk of endometrial cancer, but nulliparity might explain the increased risk of ovarian cancer. The excess of endometrial cancer was prominent among patients with normal estrogen production but progesterone deficiency (SIR = 9.4, 95% CI 5.0-16.0). The risk for ovarian cancer was similar among the total groups of treated and untreated patients (SIR = 1.7 vs. 1.6). The standardized incidence ratio for endometrial cancer was higher among the treated group than the untreated group, although not significantly. Treatment with ovulation-inducing drugs does not appear to increase the risk for ovarian cancer, but its role cannot be completely excluded.     

Lymphocytic infiltration in endometrial adenocarcinoma. Correlation with histologic differentiation, clinical staging and prognosis

Lymphocytic infiltration in endometrial adenocarcinoma was evaluated in curettings from 50 patients and correlated with histologic differentiation, clinical staging and prognosis. In spite of the absence of any significant correlation between lymphocytic infiltration and the various parameters investigated, the presence of a marked infiltrate in 46% of the cases suggests that in endometrial adenocarcinoma, as in other neoplasms, a cell-mediated immune response does indeed exist, but warrants further investigation by more refined methods.     

Adaptive Retrieval Agents: Internalizing Local Context and Scaling up to the Web

This paper discusses a novel distributed adaptive algorithm and representation used to construct populations of adaptive Web agents. These InfoSpiders browse networked information environments on-line in search of pages relevant to the user, by traversing hyperlinks in an autonomous and intelligent fashion. Each agent adapts to the spatial and temporal regularities of its local context thanks to a combination of machine learning techniques inspired by ecological models: evolutionary adaptation with local selection, reinforcement learning and selective query expansion by internalization of environmental signals, and optional relevance feedback. We evaluate the feasibility and performance of these methods in three domains: a general class of artificial graph environments, a controlled subset of the Web, and (preliminarly) the full Web. Our results suggest that InfoSpiders could take advantage of the starting points provided by search engines, based on global word statistics, and then use linkage topology to guide their search on-line. We show how this approach can complement the current state of the art, especially with respect to the scalability challenge.     

Pregnancy and delivery in a group of Israeli teenagers. A case-controlled study

BACKGROUND: Osteoarthritis is a common cause of disability in our community, particularly in the elderly. OBJECTIVE: The aim of treatment in those with osteoarthritis should be the relief of pain and maintenance of function with minimal risk of side effects. DISCUSSION: Non pharmacological therapy, tailored to the individual, should be the first line treatment of osteoarthritis patients. Paracetamol may be added if necessary. Non steroidal anti-inflammatory drugs should only be used after assessing the patient for the risk of side effects. If the risk is high then alternative therapy should be sought, or measures taken to minimise the risk.     

Emerging small-world referral networks in evolutionary labormarkets

We model a labor market that includes referral networks using an agent-based simulation. Agents maximize their employment satisfaction by allocating resources to build friendship networks and to adjust search intensity. We use a local selection evolutionary algorithm, which maintains a diverse population of strategies, to study the adaptive graph topologies resulting from the model. The evolved networks display mixtures of regularity and randomness, as in small-world networks. A second characteristic emerges in our model as time progresses: the population loses efficiency due to over competition for job referral contacts in a way similar to social dilemmas such as the tragedy of the commons. Analysis reveals that the loss of global fitness is driven by an increase in individual robustness, which allows agents to live longer by surviving job losses. The behavior of our model suggests predictions for a number of policies     

Trigger control in a high energy physics experiment

Control and diagnostic tools for the event-filtering system (trigger) of the L3 experiment at the LEP (Large Electron Positron) collider have been implemented using an expert system. The structure and behavior of the trigger electronics has been modeled using the hybrid framework (allowing both frame-based design and rule-based reasoning) of the shell NEXPERT Object, which also provides a user friendly interface. An open architecture allows the integration in the system of conventional and widespread tools, like procedural languages and database management systems, and an easy interface to the whole online environment. The project goals, base model, development tools, and present implementation and future development of the trigger control are described     

Efficient and scalable Pareto optimization by evolutionary local selection algorithms

Local selection is a simple selection scheme in evolutionary computation. Individual fitnesses are accumulated over time and compared to a fixed threshold, rather than to each other, to decide who gets to reproduce. Local selection, coupled with fitness functions stemming from the consumption of finite shared environmental resources, maintains diversity in a way similar to fitness sharing. However, it is more efficient than fitness sharing and lends itself to parallel implementations for distributed tasks. While local selection is not prone to premature convergence, it applies minimal selection pressure to the population. Local selection is, therefore, particularly suited to Pareto optimization or problem classes where diverse solutions must be covered. This paper introduces ELSA, an evolutionary algorithm employing local selection and outlines three experiments in which ELSA is applied to multiobjective problems: a multimodal graph search problem, and two Pareto optimization problems. In all these experiments, ELSA significantly outperforms other well-known evolutionary algorithms. The paper also discusses scalability, parameter dependence, and the potential distributed applications of the algorithm.     

MySpiders: Evolve Your Own Intelligent Web Crawlers

The dynamic nature of the World Wide Web makes it a challenge to find information that is both relevant and recent. Intelligent agents can complement the power of search engines to meet this challenge. We present a Web tool called MySpiders, which implements an evolutionary algorithm managing a population of adaptive crawlers who browse the Web autonomously. Each agent acts as an intelligent client on behalf of the user, driven by a user query and by textual and linkage clues in the crawled pages. Agents autonomously decide which links to follow, which clues to internalize, when to spawn offspring to focus the search near a relevant source, and when to starve. The tool is available to the public as a threaded Java applet. We discuss the development and deployment of such a system.