A robust principal component analysis

This work is concerned with robustness in Principal Component Analysis (PCA). The approach, which we adopt here, is to replace the criterion of least squares by another criterion based on a convex and sufficiently differentiable loss function ρ. Using this criterion we propose a robust estimate of the location vector and introduce an orthogonality with respect to (w.r.t.) ρ in order to define the different steps of a PCA. The influence functions of a vector mean and principal vectors are developed in order to provide method for obtaining a robust PCA. The practical procedure is based on an alternative-steps algorithm.     

SPECTACLES OF ETHNICITY: FESTIVALS AND THE COMMODIFICATION OF ETHNIC CULTURE AMONG LOUISIANA CAJUNS

Drawing on the example of the Louisiana Cajuns, an ethnic group that has been enjoying a wave of popular revival in recent years, this study suggests that changes in the perception of an ethnic identity are related to socioeconomic transformation. We identify the festival as a key aspect of the Cajun revival since the 1960s. An examination of the history, activities, and contemporary spatiotemporal organization of festivals reveals similarities to other aspects of a society of mass consumption. Niche marketing , the structuring of recreation around the modern work week, and the establishment of personal identity through the purchase of symbolically rich commodities are all embodied in present - day Cajun festivals. At the same time, the consumption of ethnic commodities is linked by the consumers with a sense of tradition and descent from a mythic past. The festivals of southwestern Louisiana are, in this sense, invented traditions and, paradoxically, a measure of the assimilation of this particular ethnic group into American culture.     

A new estimate of Ukrainian population losses during the crises of the 1930s and 1940s

Ukraine experienced two very acute demographic crises during the Soviet era: the 1933 famine and the Second World War. While different estimates of total losses have been produced previously, we have tried here to distinguish the specific impact of the crises on mortality from their impact on fertility and migration. Taking into account all existing sources of registered data and estimates, a painstaking reconstruction of annual demographic changes has been produced and complete annual life tables have been computed for the years 1926-59. Life expectancy at birth fell to a level as low as 10 years for females and 7 for males in 1933 and plateaued around 25 for females and 15 for males in the period 1941-44.     

The labour supply of married and cohabiting women in the Netherlands, 1981–1989

The analysis of labour supply involves two issues: whether individuals participate in the labour market and, if they do, for how many hours. In this article we investigate the differences in labour supply of married women and cohabiting women in the Netherlands; we try to answer the question how these differences can be explained. From this study, it can be concluded that differences between both categories of women in participation and in weekly hours worked can predominantly be explained by differences in characteristics (e.g. age, net wage rate, and age of children), than by differences in behaviour. The empirical results indicate that cohabiting women are more economically independent than married women. However for married women we found evidence that there was increased economic independence during the eighties; i.e. their weekly hours work has become less affected by the income of their partners.     

Adolescents' beliefs about marijuana use: a comparison of regular users, past users and never/occasional users

A questionnaire investigating adolescents' opinions and experiences regarding marijuana use was administered to 163 adolescents and young adults (96 boys and 67 girls) aged 13 to 20 (mean age = 16.8, s.d. = 1.5). Items referred to marijuana and other substances' dangerousness, representations regarding the positive and negative consequences of marijuana use. Responses were compared according to marijuana use status (classified into never/occasional use, current regular use and past regular use). Results show that adolescents' opinions differ according to their experience with marijuana use. Current regular users evaluate marijuana as less dangerous, but alcohol and heroin as more dangerous in comparison with never/occasional and past users. Current and past users are more likely to define marijuana as a medical drug and a plant used in agriculture, and less likely to define it as an illegal drug. Current and past users evaluate marijuana use as a way to cope with stress, to relax to a greater extent than do never/occasional users do. The latter attribute more negative consequences to marijuana use such as diminished driving ability and school performance and a pathway to hard drugs.