situations—where a user, an item, or the entire system is new—simple non-personalized recommendations often fare better. We improve the scalability and performance of a previous approach to handling cold-start situations that uses filterbots, or surrogate users that rate items based only on user or item

We have developed a method for recommending items that combines content and collaborative data under a single probabilistic framework. We benchmark our algorithm against a nave Bayes classifier on the cold-start problem, where we wish to recommend items that no one in the community has yet rated

Abstract not found

, and make recommendations for research and practice. In spite of general decreases in the prevalence of the nonmed-ical use of most legal and illegal drugs in recent years, the abuse of alcohol and other drugs during adolescence and early adulthood remains a serious public health problem (Adams, Blanken

Sequential quadratic programming (SQP) methods have proved highly effective for solving constrained optimization problems with smooth nonlinear functions in the objective and constraints. Here we consider problems with general inequality constraints (linear and nonlinear). We assume that first derivatives are available, and that the constraint gradients are sparse. We discuss

Contract No. 68-C7-0051

We have systematically made a set of precisely defined, single-gene deletions of all nonessential genes in Escherichia coli K-12. Open-reading frame coding regions were replaced with a kanamycin cassette flanked by FLP recognition target sites by using a one-step method for inactivation of chromosomal genes and primers designed to create in-frame deletions upon excision of the resistance cassette. Of 4288 genes targeted, mutants were obtained for 3985. To alleviate problems encountered in high-throughput studies, two independent mutants were saved for every deleted gene. These mutants—the ‘Keio collection’—provide a new resource not only for systematic analyses of unknown gene functions and gene regulatory networks but also for genome-wide testing of mutational effects in a common strain background, E. coli K-12 BW25113. We were unable to disrupt 303 genes, including 37 of unknown function, which are candidates for essential genes. Distribution is being handled via GenoBase

Since 1973 technological, political, regulatory, and economic forces have been changing the worldwide economy in a fashion comparable to the changes experienced during the nineteenth century Industrial Revolution. As in the nineteenth century, we are experiencing declining costs, increaing average (but decreasing marginal) productivity of labor, reduced growth rates of labor income, excess capacity, and the requirement for downsizing and exit. The last two decades indicate corporate internal control systems have failed to deal effectively with these changes, especially slow growth and the requirement for exit. The next several decades pose a major challenge for Western firms and political systems as these forces continue to work their way through the worldwide economy.

This paper discusses the prevalence of Silicon Valley–style localiza-tions of individual manufacturing industries in the United States. A model in which localized industry-specific spillovers, natural ad-vantages, and pure random chance all contribute to geographic concentration is used to develop a test for whether observed levels of concentration are greater than would be expected to arise ran-domly and to motivate new indices of geographic concentration and of coagglomeration. The proposed indices control for differ-ences in the size distribution of plants and for differences in the size of the geographic areas for which data are available. As a conse-quence, comparisons of the degree of geographic concentration across industries can be made with more confidence. Our empiri-cal results provide a strong reaffirmation of the previous wisdom in that we find almost all industries to be somewhat localized. In many industries, however, the degree of localization is slight. We explore the nature of agglomerative forces in describing patterns of concentration, the geographic scope of localization, and the coagglomeration of related industries and of industries with strong upstream-downstream ties.

Currently, most approaches to retrieving textual materials from scientific databases depend on a lexical match between words in users' requests and those in or assigned to documents in a database. Because of the tremendous diversity in the words people use to describe the same document, lexical methods are necessarily incomplete and imprecise. Using the singular value decomposition (SVD), one can take advantage of the implicit higher-order structure in the association of terms with documents by determining the SVD of large sparse term by document matrices. Terms and documents represented by 200-300 of the largest singular vectors are then matched against user queries. We call this retrieval method Latent Semantic Indexing (LSI) because the subspace represents important associative relationships between terms and documents that are not evident in individual documents. LSI is a completely automatic yet intelligent indexing method, widely applicable, and a promising way to improve users...