Data Mining Practical Machine Learning Tools and Techniques

We will review some of the major results in random graphs and some of the more challenging open problems. We will cover algorithmic and structural questions. We will touch on newer models, including those related to the WWW.

Random graphs

Twitter, a microblogging service less than three years old, commands more than 41 million users as of July 2009 and is growing fast. Twitter users tweet about any topic within the 140-character limit and follow others to receive their tweets. The goal of this paper is to study the topological characteristics of Twitter and its power as a new medium of information sharing.We have crawled the entire Twitter site and obtained 41.7 million user profiles, 1.47 billion social relations, 4,262 trending topics, and 106 million tweets. In its follower-following topology analysis we have found a non-power-law follower distribution, a short effective diameter, and low reciprocity, which all mark a deviation from known characteristics of human social networks [28]. In order to identify influentials on Twitter, we have ranked users by the number of followers and by PageRank and found two rankings to be similar. Ranking by retweets differs from the previous two rankings, indicating a gap in influence inferred from the number of followers and that from the popularity of one's tweets. We have analyzed the tweets of top trending topics and reported on their temporal behavior and user participation. We have classified the trending topics based on the active period and the tweets and show that the majority (over 85%) of topics are headline news or persistent news in nature. A closer look at retweets reveals that any retweeted tweet is to reach an average of 1,000 users no matter what the number of followers is of the original tweet. Once retweeted, a tweet gets retweeted almost instantly on next hops, signifying fast diffusion of information after the 1st retweet.To the best of our knowledge this work is the first quantitative study on the entire Twittersphere and information diffusion on it.

https://ink.library.smu.edu.sg/cgi/viewcontent.cgi?article=7104&context=sis_research

What is Twitter, a social network or a news media?

The term “Linked Data” refers to a set of best practices for publishing and connecting structured data on the Web. These best practices have been adopted by an increasing number of data providers over the last three years, leading to the creation of a global data space containing billions of assertions— the Web of Data. In this article, the authors present the concept and technical principles of Linked Data, and situate these within the broader context of related technological developments. They describe progress to date in publishing Linked Data on the Web, review applications that have been developed to exploit the Web of Data, and map out a research agenda for the Linked Data community as it moves forward.

/pdf/linked-data-the-story-so-far-2a9zvmaysa.pdf

Linked Data - the story so far

Obesity now affects one in five children in the United States. Discrimination against overweight children begins early in childhood and becomes progressively institutionalized. Because obese children tend to be taller than their nonoverweight peers, they are apt to be viewed as more mature. The inappropriate expectations that result may have an adverse effect on their socialization. Many of the cardiovascular consequences that characterize adult-onset obesity are preceded by abnormalities that begin in childhood. Hyperlipidemia, hypertension, and abnormal glucose tolerance occur with increased frequency in obese children and adolescents. The relationship of cardiovascular risk factors to visceral fat independent of total body fat remains unclear. Sleep apnea, pseudotumor cerebri, and Blount's disease represent major sources of morbidity for which rapid and sustained weight reduction is essential. Although several periods of increased risk appear in childhood, it is not clear whether obesity with onset early in childhood carries a greater risk of adult morbidity and mortality. Obesity is now the most prevalent nutritional disease of children and adolescents in the United States. Although obesity-associated morbidities occur more frequently in adults, significant consequences of obesity as well as the antecedents of adult disease occur in obese children and adolescents. In this review, I consider the adverse effects of obesity in children and adolescents and attempt to outline areas for future research. I refer to obesity as a body mass index greater than the 95th percentile for children of the same age and gender.

Health Consequences of Obesity in Youth: Childhood Predictors of Adult Disease

We describe Mink, a new web browser extension that provides a different model for integration of the live and archived web. While a user browses the live web, Mink actively queries the archives and reports other instances of the page in the archives without requiring active querying by the user. Further, by querying the archives dynamically and asynchronously, a user can view the extent to which the currently viewed page on the live web has been archived and proactively submit a request to various archives using an overlay on the live web page and a simple interface.

Mink: integrating the live and archived web viewing experience using web browsers and memento

OBJECTIVE To estimate the costs per relevant unit (pupils and meals) associated with improvements to school food and the potential economic and health gains that may result. DESIGN Calculation of costs per relevant unit (pupils and meals) based on (i) Department for Education expenditure to support improvements in school food, 2005–2011 and (ii) measures of the changes in the number of pupils taking school lunch and the number of meals served over the same time period; plus examples of the use of linked data to predict longer-term economic and health outcomes of healthier eating at school. SETTING England. SUBJECTS Local authorities, government departments and non-departmental public bodies. RESULTS Analysis of investment over a 6-year period indicates that costs of setting up and maintaining a change organization such as the School Food Trust were low in relation to short-term benefits in nutrition and behaviour. Models that predict long-terms gains to the exchequer and to quality-adjusted life years need further elaboration. CONCLUSIONS Modest levels of government investment in the delivery and promotion of healthier school food is likely to yield both short-term and long-term benefits in relation to nutrition, learning, economics and health.

/pdf/school-food-cost-benefits-england-50brngzcqr.pdf

School food cost-benefits: England.

Buckets are an aggregative, intelligent construct for publishing in digital libraries. The goal of research projects is to produce information. This information is often instantiated in several forms, differentiated by semantic types (report, software, video, datasets, etc.). A given semantic type can be further differentiated by syntactic representations as well (PostScript version, PDF version, Word version, etc.). Although the information was created together and subtle relationships can exist between them, different semantic instantiations are generally segregated along currently obsolete media boundaries. Reports are placed in report archives, software might go into a software archive, but most of the data and supporting materials are likely to be kept in informal personal archives or discarded altogether. Buckets provide an archive-independent container construct in which all related semantic and syntactic data types and objects can be logically grouped together, archived, and manipulated as a single object. Furthermore, buckets are active archival objects and can communicate with each other, people, or arbitrary network services.

/pdf/buckets-aggregative-intelligent-agents-for-publishing-2xom6zb2gn.pdf

Buckets: Aggregative, Intelligent Agents for Publishing

This paper discusses the requirements of current and emerging applications based on the Open Archives Initiative (OAI) and emphasizes the need for a common infrastructure to support them. Inspired by HTTP proxy, cache, gateway and web service concepts, a design for a scalable and reliable infrastructure that aims at satisfying these requirements is presented. Moreover it is shown how various applications can exploit the services included in the proposed infrastructure. The paper concludes by discussing the current status of several prototype implementations.

https://digitalcommons.odu.edu/cgi/viewcontent.cgi?article=1010&context=computerscience_fac_pubs

A Scalable Architecture for Harvest-Based Digital Libraries - The ODU/Southampton Experiments

An archive profile is a high-level summary of a web archive’s holdings that can be used for routing Memento queries to the appropriate archives. It can be created by generating summaries from the CDX files (index of web archives) which we explored in an earlier work. However, requiring archives to update their profiles periodically is difficult. Alternative means to discover the holdings of an archive involve sampling based approaches such as fulltext keyword searching to learn the URIs present in the response or looking up for a sample set of URIs and see which of those are present in the archive. It is the fulltext search based discovery and profiling that is the scope of this paper. We developed the Random Searcher Model (RSM) to discover the holdings of an archive by a random search walk. We measured the search cost of discovering certain percentages of the archive holdings for various profiling policies under different RSM configurations. We can make routing decisions of 80 % of the requests correctly while maintaining about 0.9 recall by discovering only 10 % of the archive holdings and generating a profile that costs less than 1 % of the complete knowledge profile.

Michael L. Nelson

Papers

Mink: integrating the live and archived web viewing experience using web browsers and memento

School food cost-benefits: England.

Buckets: Aggregative, Intelligent Agents for Publishing

A Scalable Architecture for Harvest-Based Digital Libraries - The ODU/Southampton Experiments

Web Archive Profiling Through Fulltext Search