Clustering data is not an easy problem in general, and is compounded for a massive dataset. Restricting attention to a sample from the data ignores minority groups and hence compromises on the available riches. This paper develops, under Gaussian assumptions, a multistage sequential algorithm

www.nber.org/~jwolfers The “Easterlin Paradox ” suggests that there is no link between the level of economic development of a society and average levels of happiness. We return to Easterlin’s question: “Will raising the incomes of all increase the happiness of all? ” and analyze multiple rich

Recent growth of the geo-spatial information on the web has made it possible to easily access a wide variety of spatial data. By integrating these spatial datasets, one can support a rich set of queries that could not have been answered given any of these sets in isolation. However, accurately

introduce the Caltech Pedestrian Dataset, which is two orders of magnitude larger than existing datasets. The dataset contains richly annotated video, recorded from a moving vehicle, with challenging images of low resolution and frequently occluded people. We propose improved evaluation metrics

Do students benefit from compulsory schooling? In an important article, Oreopoulos (2006) studied the 1947 British compulsory schooling law change and found large returns to schooling of about 15 % using the General Household Survey (GHS). We reanalyse this dataset and find much smaller returns

High-throughput biological screens are yielding ever-growing streams of information about multiple aspects of cellular activity. As more and more categories of datasets come online, there is a corresponding multitude of ways in which inferences can be chained across them, motivating the need

Distributing labeling tasks among hundreds or thousands of annotators is an increasingly important method for annotating large datasets. We present a method for estimating the underlying value (e.g. the class) of each image from (noisy) annotations provided by multiple annotators. Our method

In this work we identify the structure inherent in daily human behavior with models that can accurately analyze, predict and cluster multimodal data from individuals and groups. We represent this structure by the principal components of the complete behavioral dataset, a set of characteristic

hard-to-measure characteristics. By applying within twin techniques using an unusually rich dataset from Norway, we examine the effects of birth weight on both short-run and long-run outcomes for the same cohorts. We find that birth weight does matter; despite short-run twin fixed effects estimates

Accurate recognition and tracking of human activities is an important goal of ubiquitous computing. Recent advances in the development of multi-modal wearable sensors enable us to gather rich datasets of human activities. However, the problem of automatically identifying the most useful features