Search (29 results, page 1 of 2)

0.02067415 = product of:
  0.0413483 = sum of:
    0.0413483 = sum of:
      0.010148063 = weight(_text_:a in 2565) [ClassicSimilarity], result of:
        0.010148063 = score(doc=2565,freq=18.0), product of:
          0.053105544 = queryWeight, product of:
            1.153047 = idf(docFreq=37942, maxDocs=44218)
            0.046056706 = queryNorm
          0.19109234 = fieldWeight in 2565, product of:
            4.2426405 = tf(freq=18.0), with freq of:
              18.0 = termFreq=18.0
            1.153047 = idf(docFreq=37942, maxDocs=44218)
            0.0390625 = fieldNorm(doc=2565)
      0.03120024 = weight(_text_:22 in 2565) [ClassicSimilarity], result of:
        0.03120024 = score(doc=2565,freq=2.0), product of:
          0.16128273 = queryWeight, product of:
            3.5018296 = idf(docFreq=3622, maxDocs=44218)
            0.046056706 = queryNorm
          0.19345059 = fieldWeight in 2565, product of:
            1.4142135 = tf(freq=2.0), with freq of:
              2.0 = termFreq=2.0
            3.5018296 = idf(docFreq=3622, maxDocs=44218)
            0.0390625 = fieldNorm(doc=2565)
  0.5 = coord(1/2)

Abstract

This paper introduces a family of link-based ranking algorithms that propagate page importance through links. In these algorithms there is a damping function that decreases with distance, so a direct link implies more endorsement than a link through a long path. PageRank is the most widely known ranking function of this family. The main objective of this paper is to determine whether this family of ranking techniques has some interest per se, and how different choices for the damping function impact on rank quality and on convergence speed. Even though our results suggest that PageRank can be approximated with other simpler forms of rankings that may be computed more efficiently, our focus is of more speculative nature, in that it aims at separating the kernel of PageRank, that is, link-based importance propagation, from the way propagation decays over paths. We focus on three damping functions, having linear, exponential, and hyperbolic decay on the lengths of the paths. The exponential decay corresponds to PageRank, and the other functions are new. Our presentation includes algorithms, analysis, comparisons and experiments that study their behavior under different parameters in real Web graph data. Among other results, we show how to calculate a linear approximation that induces a page ordering that is almost identical to PageRank's using a fixed small number of iterations; comparisons were performed using Kendall's tau on large domain datasets.

Date

16. 1.2016 10:22:28

Type

a

0.020383961 = product of:
  0.040767923 = sum of:
    0.040767923 = sum of:
      0.009567685 = weight(_text_:a in 2564) [ClassicSimilarity], result of:
        0.009567685 = score(doc=2564,freq=16.0), product of:
          0.053105544 = queryWeight, product of:
            1.153047 = idf(docFreq=37942, maxDocs=44218)
            0.046056706 = queryNorm
          0.18016359 = fieldWeight in 2564, product of:
            4.0 = tf(freq=16.0), with freq of:
              16.0 = termFreq=16.0
            1.153047 = idf(docFreq=37942, maxDocs=44218)
            0.0390625 = fieldNorm(doc=2564)
      0.03120024 = weight(_text_:22 in 2564) [ClassicSimilarity], result of:
        0.03120024 = score(doc=2564,freq=2.0), product of:
          0.16128273 = queryWeight, product of:
            3.5018296 = idf(docFreq=3622, maxDocs=44218)
            0.046056706 = queryNorm
          0.19345059 = fieldWeight in 2564, product of:
            1.4142135 = tf(freq=2.0), with freq of:
              2.0 = termFreq=2.0
            3.5018296 = idf(docFreq=3622, maxDocs=44218)
            0.0390625 = fieldNorm(doc=2564)
  0.5 = coord(1/2)

Abstract

PageRank is defined as the stationary state of a Markov chain. The chain is obtained by perturbing the transition matrix induced by a web graph with a damping factor alpha that spreads uniformly part of the rank. The choice of alpha is eminently empirical, and in most cases the original suggestion alpha=0.85 by Brin and Page is still used. Recently, however, the behaviour of PageRank with respect to changes in alpha was discovered to be useful in link-spam detection. Moreover, an analytical justification of the value chosen for alpha is still missing. In this paper, we give the first mathematical analysis of PageRank when alpha changes. In particular, we show that, contrarily to popular belief, for real-world graphs values of alpha close to 1 do not give a more meaningful ranking. Then, we give closed-form formulae for PageRank derivatives of any order, and an extension of the Power Method that approximates them with convergence O(t**k*alpha**t) for the k-th derivative. Finally, we show a tight connection between iterated computation and analytical behaviour by proving that the k-th iteration of the Power Method gives exactly the PageRank value obtained using a Maclaurin polynomial of degree k. The latter result paves the way towards the application of analytical methods to the study of PageRank.

Date

16. 1.2016 10:22:28

Type

a

0.0028703054 = product of:
  0.005740611 = sum of:
    0.005740611 = product of:
      0.011481222 = sum of:
        0.011481222 = weight(_text_:a in 4704) [ClassicSimilarity], result of:
          0.011481222 = score(doc=4704,freq=16.0), product of:
            0.053105544 = queryWeight, product of:
              1.153047 = idf(docFreq=37942, maxDocs=44218)
              0.046056706 = queryNorm
            0.2161963 = fieldWeight in 4704, product of:
              4.0 = tf(freq=16.0), with freq of:
                16.0 = termFreq=16.0
              1.153047 = idf(docFreq=37942, maxDocs=44218)
              0.046875 = fieldNorm(doc=4704)
      0.5 = coord(1/2)
  0.5 = coord(1/2)

Abstract

Swoogle is a crawler-based indexing and retrieval system for the Semantic Web, i.e., for Web documents in RDF or OWL. It extracts metadata for each discovered document, and computes relations between documents. Discovered documents are also indexed by an information retrieval system which can use either character N-Gram or URIrefs as keywords to find relevant documents and to compute the similarity among a set of documents. One of the interesting properties we compute is rank, a measure of the importance of a Semantic Web document.

Content

Vgl. unter: http://www.dblab.ntua.gr/~bikakis/LD/5.pdf Vgl. auch: http://swoogle.umbc.edu/. Vgl. auch: http://ebiquity.umbc.edu/paper/html/id/183/. Vgl. auch: Radhakrishnan, A.: Swoogle : An Engine for the Semantic Web unter: http://www.searchenginejournal.com/swoogle-an-engine-for-the-semantic-web/5469/.

Type

a

0.00270615 = product of:
  0.0054123 = sum of:
    0.0054123 = product of:
      0.0108246 = sum of:
        0.0108246 = weight(_text_:a in 910) [ClassicSimilarity], result of:
          0.0108246 = score(doc=910,freq=2.0), product of:
            0.053105544 = queryWeight, product of:
              1.153047 = idf(docFreq=37942, maxDocs=44218)
              0.046056706 = queryNorm
            0.20383182 = fieldWeight in 910, product of:
              1.4142135 = tf(freq=2.0), with freq of:
                2.0 = termFreq=2.0
              1.153047 = idf(docFreq=37942, maxDocs=44218)
              0.125 = fieldNorm(doc=910)
      0.5 = coord(1/2)
  0.5 = coord(1/2)

0.00270615 = product of:
  0.0054123 = sum of:
    0.0054123 = product of:
      0.0108246 = sum of:
        0.0108246 = weight(_text_:a in 384) [ClassicSimilarity], result of:
          0.0108246 = score(doc=384,freq=2.0), product of:
            0.053105544 = queryWeight, product of:
              1.153047 = idf(docFreq=37942, maxDocs=44218)
              0.046056706 = queryNorm
            0.20383182 = fieldWeight in 384, product of:
              1.4142135 = tf(freq=2.0), with freq of:
                2.0 = termFreq=2.0
              1.153047 = idf(docFreq=37942, maxDocs=44218)
              0.125 = fieldNorm(doc=384)
      0.5 = coord(1/2)
  0.5 = coord(1/2)

0.00270615 = product of:
  0.0054123 = sum of:
    0.0054123 = product of:
      0.0108246 = sum of:
        0.0108246 = weight(_text_:a in 5595) [ClassicSimilarity], result of:
          0.0108246 = score(doc=5595,freq=8.0), product of:
            0.053105544 = queryWeight, product of:
              1.153047 = idf(docFreq=37942, maxDocs=44218)
              0.046056706 = queryNorm
            0.20383182 = fieldWeight in 5595, product of:
              2.828427 = tf(freq=8.0), with freq of:
                8.0 = termFreq=8.0
              1.153047 = idf(docFreq=37942, maxDocs=44218)
              0.0625 = fieldNorm(doc=5595)
      0.5 = coord(1/2)
  0.5 = coord(1/2)

Abstract

The study of the web as a graph is not only fascinating in its own right, but also yields valuable insight into web algorithms for crawling, searching and community discovery, and the sociological phenomena which characterize its evolution. We report on experiments on local and global properties of the web graph using two Altavista crawls each with over 200M pages and 1.5 billion links. Our study indicates that the macroscopic structure of the web is considerably more intricate than suggested by earlier experiments on a smaller scale

0.00270615 = product of:
  0.0054123 = sum of:
    0.0054123 = product of:
      0.0108246 = sum of:
        0.0108246 = weight(_text_:a in 706) [ClassicSimilarity], result of:
          0.0108246 = score(doc=706,freq=2.0), product of:
            0.053105544 = queryWeight, product of:
              1.153047 = idf(docFreq=37942, maxDocs=44218)
              0.046056706 = queryNorm
            0.20383182 = fieldWeight in 706, product of:
              1.4142135 = tf(freq=2.0), with freq of:
                2.0 = termFreq=2.0
              1.153047 = idf(docFreq=37942, maxDocs=44218)
              0.125 = fieldNorm(doc=706)
      0.5 = coord(1/2)
  0.5 = coord(1/2)

0.0026473717 = product of:
  0.0052947435 = sum of:
    0.0052947435 = product of:
      0.010589487 = sum of:
        0.010589487 = weight(_text_:a in 4838) [ClassicSimilarity], result of:
          0.010589487 = score(doc=4838,freq=10.0), product of:
            0.053105544 = queryWeight, product of:
              1.153047 = idf(docFreq=37942, maxDocs=44218)
              0.046056706 = queryNorm
            0.19940455 = fieldWeight in 4838, product of:
              3.1622777 = tf(freq=10.0), with freq of:
                10.0 = termFreq=10.0
              1.153047 = idf(docFreq=37942, maxDocs=44218)
              0.0546875 = fieldNorm(doc=4838)
      0.5 = coord(1/2)
  0.5 = coord(1/2)

Abstract

DAFFODIL is a search system for digital libraries aiming at strategic support during the information search process. From a user point of view this strategic support is mainly implemented by high-level search functions, so-called stratagems, which provide functionality beyond today's digital libraries. Through the tight integration of stratagems and with the federation of heterogeneous digital libraries, DAFFODIL reaches high effects of synergy for information and services. These effects provide high-quality metadata for the searcher through an intuitively controllable user interface. The implementation of stratagems follows a tool-based model.

Type

a

0.0024857575 = product of:
  0.004971515 = sum of:
    0.004971515 = product of:
      0.00994303 = sum of:
        0.00994303 = weight(_text_:a in 1215) [ClassicSimilarity], result of:
          0.00994303 = score(doc=1215,freq=12.0), product of:
            0.053105544 = queryWeight, product of:
              1.153047 = idf(docFreq=37942, maxDocs=44218)
              0.046056706 = queryNorm
            0.18723148 = fieldWeight in 1215, product of:
              3.4641016 = tf(freq=12.0), with freq of:
                12.0 = termFreq=12.0
              1.153047 = idf(docFreq=37942, maxDocs=44218)
              0.046875 = fieldNorm(doc=1215)
      0.5 = coord(1/2)
  0.5 = coord(1/2)

Abstract

Directed Query Engines, an emerging class of search engine specifically designed to access distributed resources on the deep web, offer the opportunity to create inexpensive digital libraries. Already, one such engine, Distributed Explorer, has been used to select and assemble high quality information resources and incorporate them into publicly available systems for the physical sciences. By nesting Directed Query Engines so that one query launches several other engines in a cascading fashion, enormous virtual collections may soon be assembled to form a comprehensive information infrastructure for the physical sciences. Once a Directed Query Engine has been configured for a set of information resources, distributed alerts tools can provide patrons with personalized, profile-based notices of recent additions to any of the selected resources. Due to the potentially enormous size and scope of Directed Query Engine applications, consideration must be given to issues surrounding the representation of large quantities of information from multiple, heterogeneous sources.

Type

a

0.002440756 = product of:
  0.004881512 = sum of:
    0.004881512 = product of:
      0.009763024 = sum of:
        0.009763024 = weight(_text_:a in 93) [ClassicSimilarity], result of:
          0.009763024 = score(doc=93,freq=34.0), product of:
            0.053105544 = queryWeight, product of:
              1.153047 = idf(docFreq=37942, maxDocs=44218)
              0.046056706 = queryNorm
            0.1838419 = fieldWeight in 93, product of:
              5.8309517 = tf(freq=34.0), with freq of:
                34.0 = termFreq=34.0
              1.153047 = idf(docFreq=37942, maxDocs=44218)
              0.02734375 = fieldNorm(doc=93)
      0.5 = coord(1/2)
  0.5 = coord(1/2)

Abstract

Imagine a library containing 25 billion documents but with no centralized organization and no librarians. In addition, anyone may add a document at any time without telling anyone. You may feel sure that one of the documents contained in the collection has a piece of information that is vitally important to you, and, being impatient like most of us, you'd like to find it in a matter of seconds. How would you go about doing it? Posed in this way, the problem seems impossible. Yet this description is not too different from the World Wide Web, a huge, highly-disorganized collection of documents in many different formats. Of course, we're all familiar with search engines (perhaps you found this article using one) so we know that there is a solution. This article will describe Google's PageRank algorithm and how it returns pages from the web's collection of 25 billion documents that match search criteria so well that "google" has become a widely used verb. Most search engines, including Google, continually run an army of computer programs that retrieve pages from the web, index the words in each document, and store this information in an efficient format. Each time a user asks for a web search using a search phrase, such as "search engine," the search engine determines all the pages on the web that contains the words in the search phrase. (Perhaps additional information such as the distance between the words "search" and "engine" will be noted as well.) Here is the problem: Google now claims to index 25 billion pages. Roughly 95% of the text in web pages is composed from a mere 10,000 words. This means that, for most searches, there will be a huge number of pages containing the words in the search phrase. What is needed is a means of ranking the importance of the pages that fit the search criteria so that the pages can be sorted with the most important pages at the top of the list. One way to determine the importance of pages is to use a human-generated ranking. For instance, you may have seen pages that consist mainly of a large number of links to other resources in a particular area of interest. Assuming the person maintaining this page is reliable, the pages referenced are likely to be useful. Of course, the list may quickly fall out of date, and the person maintaining the list may miss some important pages, either unintentionally or as a result of an unstated bias. Google's PageRank algorithm assesses the importance of web pages without human evaluation of the content. In fact, Google feels that the value of its service is largely in its ability to provide unbiased results to search queries; Google claims, "the heart of our software is PageRank." As we'll see, the trick is to ask the web itself to rank the importance of pages.

0.0023678814 = product of:
  0.0047357627 = sum of:
    0.0047357627 = product of:
      0.009471525 = sum of:
        0.009471525 = weight(_text_:a in 3961) [ClassicSimilarity], result of:
          0.009471525 = score(doc=3961,freq=2.0), product of:
            0.053105544 = queryWeight, product of:
              1.153047 = idf(docFreq=37942, maxDocs=44218)
              0.046056706 = queryNorm
            0.17835285 = fieldWeight in 3961, product of:
              1.4142135 = tf(freq=2.0), with freq of:
                2.0 = termFreq=2.0
              1.153047 = idf(docFreq=37942, maxDocs=44218)
              0.109375 = fieldNorm(doc=3961)
      0.5 = coord(1/2)
  0.5 = coord(1/2)

Type

a

0.0023678814 = product of:
  0.0047357627 = sum of:
    0.0047357627 = product of:
      0.009471525 = sum of:
        0.009471525 = weight(_text_:a in 1161) [ClassicSimilarity], result of:
          0.009471525 = score(doc=1161,freq=8.0), product of:
            0.053105544 = queryWeight, product of:
              1.153047 = idf(docFreq=37942, maxDocs=44218)
              0.046056706 = queryNorm
            0.17835285 = fieldWeight in 1161, product of:
              2.828427 = tf(freq=8.0), with freq of:
                8.0 = termFreq=8.0
              1.153047 = idf(docFreq=37942, maxDocs=44218)
              0.0546875 = fieldNorm(doc=1161)
      0.5 = coord(1/2)
  0.5 = coord(1/2)

Abstract

With the development of the World Wide Web, the "information search" has grown to be a significant business sector of a global, competitive and commercial market. Powerful players have entered this market, such as commercial internet search engines, information portals, multinational publishers and online content integrators. Will Google, Yahoo or Microsoft be the only portals to global knowledge in 2010? If libraries do not want to become marginalized in a key area of their traditional services, they need to acknowledge the challenges that come with the globalisation of scholarly information, the existence and further growth of the academic internet

Type

a

0.002269176 = product of:
  0.004538352 = sum of:
    0.004538352 = product of:
      0.009076704 = sum of:
        0.009076704 = weight(_text_:a in 940) [ClassicSimilarity], result of:
          0.009076704 = score(doc=940,freq=10.0), product of:
            0.053105544 = queryWeight, product of:
              1.153047 = idf(docFreq=37942, maxDocs=44218)
              0.046056706 = queryNorm
            0.1709182 = fieldWeight in 940, product of:
              3.1622777 = tf(freq=10.0), with freq of:
                10.0 = termFreq=10.0
              1.153047 = idf(docFreq=37942, maxDocs=44218)
              0.046875 = fieldNorm(doc=940)
      0.5 = coord(1/2)
  0.5 = coord(1/2)

Abstract

Traditional on-line search services such as Dialog, DataStar and Lexis provide a wide range of search features (boolean and proximity operators, truncation, etc). This paper discusses the use of these features for effective searching, and argues that these features are required, regardless of advances in search engine technology. The literature on on-line searching is reviewed, identifying features that searchers find desirable for effective searching. A selective survey of current digital libraries available on the Web was undertaken, identifying which search features are present. The survey indicates that current digital libraries do not implement a wide range of search features. For instance: under half of the examples included controlled vocabulary, under half had proximity searching, only one enabled browsing of term indexes, and none of the digital libraries enable searchers to refine an initial search. Suggestions are made for enhancing the search effectiveness of digital libraries; for instance, by providing a full range of search operators, enabling browsing of search terms, enhancement of records with controlled vocabulary, enabling the refining of initial searches, etc.

Type

a

0.002269176 = product of:
  0.004538352 = sum of:
    0.004538352 = product of:
      0.009076704 = sum of:
        0.009076704 = weight(_text_:a in 3072) [ClassicSimilarity], result of:
          0.009076704 = score(doc=3072,freq=10.0), product of:
            0.053105544 = queryWeight, product of:
              1.153047 = idf(docFreq=37942, maxDocs=44218)
              0.046056706 = queryNorm
            0.1709182 = fieldWeight in 3072, product of:
              3.1622777 = tf(freq=10.0), with freq of:
                10.0 = termFreq=10.0
              1.153047 = idf(docFreq=37942, maxDocs=44218)
              0.046875 = fieldNorm(doc=3072)
      0.5 = coord(1/2)
  0.5 = coord(1/2)

Abstract

We present an analysis of the Yahoo Buzz Index over a period of 45 weeks. Our key findings are that: (1) It is most common for a search term to show up on the index for one week, followed by two weeks, three weeks, etc. Only two terms persist for all 45 weeks studied - Britney Spears and Jennifer Lopez. Search term longevity follows a power-law distribution or a winner-take-all structure; (2) Most search terms focus on entertainment. Search terms related to serious topics are found less often. The Buzz Index does not necessarily follow the "news cycle"; and, (3) We provide two ways to determine "star power" of various search terms - one that emphasizes staying power on the Index and another that emphasizes rank. In general, the methods lead to dramatically different results. Britney Spears performs well in both methods. We conclude that the data available on the Index is symptomatic of a celebrity-crazed, entertainment-centered culture.

0.002269176 = product of:
  0.004538352 = sum of:
    0.004538352 = product of:
      0.009076704 = sum of:
        0.009076704 = weight(_text_:a in 852) [ClassicSimilarity], result of:
          0.009076704 = score(doc=852,freq=10.0), product of:
            0.053105544 = queryWeight, product of:
              1.153047 = idf(docFreq=37942, maxDocs=44218)
              0.046056706 = queryNorm
            0.1709182 = fieldWeight in 852, product of:
              3.1622777 = tf(freq=10.0), with freq of:
                10.0 = termFreq=10.0
              1.153047 = idf(docFreq=37942, maxDocs=44218)
              0.046875 = fieldNorm(doc=852)
      0.5 = coord(1/2)
  0.5 = coord(1/2)

Abstract

Nutch is an open-source Web search engine that can be used at global, local, and even personal scale. Its initial design goal was to enable a transparent alternative for global Web search in the public interest - one of its signature features is the ability to "explain" its result rankings. Recent work has emphasized how it can also be used for intranets; by local communities with richer data models, such as the Creative Commons metadata-enabled search for licensed content; on a personal scale to index a user's files, email, and web-surfing history; and we also report on several other research projects built on Nutch. In this paper, we present how the architecture of the Nutch system enables it to be more flexible and scalable than other comparable systems today.

0.002269176 = product of:
  0.004538352 = sum of:
    0.004538352 = product of:
      0.009076704 = sum of:
        0.009076704 = weight(_text_:a in 2809) [ClassicSimilarity], result of:
          0.009076704 = score(doc=2809,freq=10.0), product of:
            0.053105544 = queryWeight, product of:
              1.153047 = idf(docFreq=37942, maxDocs=44218)
              0.046056706 = queryNorm
            0.1709182 = fieldWeight in 2809, product of:
              3.1622777 = tf(freq=10.0), with freq of:
                10.0 = termFreq=10.0
              1.153047 = idf(docFreq=37942, maxDocs=44218)
              0.046875 = fieldNorm(doc=2809)
      0.5 = coord(1/2)
  0.5 = coord(1/2)

Abstract

The hbz (Academic Library Center, Cologne) has a strong focus on search engine applications: Beyond the projected integration of respective technologies into the new release of the Digital Library portal solution (DigiBib6), vascoda background services also apply and take advantage of search engine technology. Experience since 2003 has given proof that building and updating of search engine indexes involves a vast amount of resources. The use of search engine federations, however, pledges major improvements: The total amount of data records held in linked indexes can be almost unlimited but also allow for a joint output of all hits retrieved. A federation also comes with excellent response times - hits retrieved can also refer to or link into the original system's layout. Nonetheless, the major challenge these days is different search engine technologies, e.g. Lucene and FAST, the variations in terms of ranking, and the implementation or non-implementation of so-called drill-downs. The lecture is designed to give a brief insight into the hbz search engine workshop with an introduction to the special project state of play.

0.0021393995 = product of:
  0.004278799 = sum of:
    0.004278799 = product of:
      0.008557598 = sum of:
        0.008557598 = weight(_text_:a in 1555) [ClassicSimilarity], result of:
          0.008557598 = score(doc=1555,freq=80.0), product of:
            0.053105544 = queryWeight, product of:
              1.153047 = idf(docFreq=37942, maxDocs=44218)
              0.046056706 = queryNorm
            0.16114321 = fieldWeight in 1555, product of:
              8.944272 = tf(freq=80.0), with freq of:
                80.0 = termFreq=80.0
              1.153047 = idf(docFreq=37942, maxDocs=44218)
              0.015625 = fieldNorm(doc=1555)
      0.5 = coord(1/2)
  0.5 = coord(1/2)

Content

"Introduction Yahoo! is the undisputed king of the Web directories, providing one of the key information navigation tools on the Internet. It has maintained its popularity over many Internet-years as the most visited Web site, against intense competition. This is because it does a good job of shifting, cataloguing and organising the Web [1] . But what would a map of Yahoo!'s hierarchical classification of the Web look like? Would an interactive map of Yahoo!, rather than the conventional listing of sites, be more useful as navigational tool? We can get some idea what a map of Yahoo! might be like by taking a look at ET-Map, a prototype developed by Hsinchun Chen and colleagues in the Artificial Intelligence Lab [2] at the University of Arizona. ET-Map was developed in 1995 as part of innovative research in automatic Internet homepage categorization and it charts a large chunk of Yahoo!, from the entertainment section representing some 110,000 different Web links. The map is a two-dimensional, multi-layered category map; its aim is to provide an intuitive visual information browsing tool. ET-Map can be browsed interactively, explored and queried, using the familiar point-and-click navigation style of the Web to find information of interest.
The View From Above Browsing for a particular piece on information on the Web can often feel like being stuck in an unfamiliar part of town walking around at street level looking for a particular store. You know the store is around there somewhere, but your viewpoint at ground level is constrained. What you really want is to get above the streets, hovering half a mile or so up in the air, to see the whole neighbourhood. This kind of birds-eye view function has been memorably described by David D. Clark, Senior Research Scientist at MIT's Laboratory for Computer Science and the Chairman of the Invisible Worlds Protocol Advisory Board, as the missing "up button" on the browser [3] . ET-Map is a nice example of a prototype for Clark's "up-button" view of an information space. The goal of information maps, like ET-Map, is to provide the browser with a sense of the lie of the information landscape, what is where, the location of clusters and hotspots, what is related to what. Ideally, this 'big-picture' all-in-one visual summary needs to fit on a single standard computer screen. ET-Map is one of my favourite examples, but there are many other interesting information maps being developed by other researchers and companies (see inset at the bottom of this page). How does ET-Map work? Here is a sequence of screenshots of a typical browsing session with ET-Map, which ends with access to Web pages on jazz musician Miles Davis. You can also tryout ET-Map for yourself, using a fully working demo on the AI Lab's website [4] . We begin with the top-level map showing forty odd broad entertainment 'subject regions' represented by regularly shaped tiles. Each tile is a visual summary of a group of Web pages with similar content. These tiles are shaded different colours to differentiate them, while labels identify the subject of the tile and the number in brackets telling you how many individual Web page links it contains. ET-Map uses two important, but common-sense, spatial concepts in its organisation and representation of the Web. Firstly, the 'subject regions' size is directly related to the number of Web pages in that category. For example, the 'MUSIC' subject area contains over 11,000 pages and so has a much larger area than the neighbouring area of 'LIVE' which only has 4,300 odd pages. This is intuitively meaningful, as the largest tiles are visually more prominent on the map and are likely to be more significant as they contain the most links. In addition, a second spatial concept, that of neighbourhood proximity, is applied so 'subject regions' closely related in term of content are plotted close to each other on the map. For example, 'FILM' and 'YEAR'S OSCARS', at the bottom left, are neighbours in both semantic and spatial space. This make senses as many things in the real-world are ordered in this way, with things that are alike being spatially close together (e.g. layout of goods in a store, or books in a library). Importantly, ET-Map is also a multi-layer map, with sub-maps showing greater informational resolution through a finer degree of categorization. So for any subject region that contains more than two hundred Web pages, a second-level map, with more detailed categories is generated. This subdivision of information space is repeated down the hierarchy as far as necessary. In the example, the user selected the 'MUSIC' subject region which, not surprisingly, contained many thousands of pages. A second-level map with numerous different music categories is then presented to the user. Delving deeper, the user wants to learn more about jazz music, so clicking on the 'JAZZ' tile leads to a third-level map, a fine-grained map of jazz related Web pages. Finally, selecting the 'MILES DAVIS' subject region leads to more a conventional looking ranking of pages from which the user selects one to download.
ET-Map was created using a sophisticated AI technique called Kohonen self-organizing map, a neural network approach that has been used for automatic analysis and classification of semantic content of text documents like Web pages. I do not pretend to fully understand how this technique works; I tend to think of it as a clever 'black-box' that group together things that are alike [5] . It is a real challenge to automatically classify pages from a very heterogeneous information collection like the Web into categories that will match the conceptions of a typical user. Directories like Yahoo! tend to rely on the skill of human editors to achieve this. ET-Map is an interesting prototype that I think highlights well the potential for a map-based approach to Web browsing. I am surprised none of the major search engines or directories have introduced the option of mapping results. Although, I am sure many are working on ideas. People certainly need all the help they get, as Web growth shows no sign of slowing. Just last month it was reported that the Web had surpassed one billion indexable pages [6].
Information Maps There are many other fascinating examples that employ two dimensional interactive maps to provide a 'birds-eye' view of information. They use various underlying techniques of textual analysis and clustering to turn the mass of information into a useful summary map (see "Mining in Textual Mountains" in Mappa.Mundi Magazine). In terms of visual representations they can be divided into two groups, those that generate smooth surfaces and those that produce regular, tiled maps. Unfortunately, we don't have space to examine them in detail, but they are well worth spending some time exploring. I will be covering some of them in future columns.

0.0021393995 = product of:
  0.004278799 = sum of:
    0.004278799 = product of:
      0.008557598 = sum of:
        0.008557598 = weight(_text_:a in 4709) [ClassicSimilarity], result of:
          0.008557598 = score(doc=4709,freq=20.0), product of:
            0.053105544 = queryWeight, product of:
              1.153047 = idf(docFreq=37942, maxDocs=44218)
              0.046056706 = queryNorm
            0.16114321 = fieldWeight in 4709, product of:
              4.472136 = tf(freq=20.0), with freq of:
                20.0 = termFreq=20.0
              1.153047 = idf(docFreq=37942, maxDocs=44218)
              0.03125 = fieldNorm(doc=4709)
      0.5 = coord(1/2)
  0.5 = coord(1/2)

Content

"Swoogle, the Semantic web search engine, is a research project carried out by the ebiquity research group in the Computer Science and Electrical Engineering Department at the University of Maryland. It's an engine tailored towards finding documents on the semantic web. The whole research paper is available here. Semantic web is touted as the next generation of online content representation where the web documents are represented in a language that is not only easy for humans but is machine readable (easing the integration of data as never thought possible) as well. And the main elements of the semantic web include data model description formats such as Resource Description Framework (RDF), a variety of data interchange formats (e.g. RDF/XML, Turtle, N-Triples), and notations such as RDF Schema (RDFS), the Web Ontology Language (OWL), all of which are intended to provide a formal description of concepts, terms, and relationships within a given knowledge domain (Wikipedia). And Swoogle is an attempt to mine and index this new set of web documents. The engine performs crawling of semantic documents like most web search engines and the search is available as web service too. The engine is primarily written in Java with the PHP used for the front-end and MySQL for database. Swoogle is capable of searching over 10,000 ontologies and indexes more that 1.3 million web documents. It also computes the importance of a Semantic Web document. The techniques used for indexing are the more google-type page ranking and also mining the documents for inter-relationships that are the basis for the semantic web. For more information on how the RDF framework can be used to relate documents, read the link here. Being a research project, and with a non-commercial motive, there is not much hype around Swoogle. However, the approach to indexing of Semantic web documents is an approach that most engines will have to take at some point of time. When the Internet debuted, there were no specific engines available for indexing or searching. The Search domain only picked up as more and more content became available. One fundamental question that I've always wondered about it is - provided that the search engines return very relevant results for a query - how to ascertain that the documents are indeed the most relevant ones available. There is always an inherent delay in indexing of document. Its here that the new semantic documents search engines can close delay. Experimenting with the concept of Search in the semantic web can only bore well for the future of search technology."

0.001913537 = product of:
  0.003827074 = sum of:
    0.003827074 = product of:
      0.007654148 = sum of:
        0.007654148 = weight(_text_:a in 1196) [ClassicSimilarity], result of:
          0.007654148 = score(doc=1196,freq=16.0), product of:
            0.053105544 = queryWeight, product of:
              1.153047 = idf(docFreq=37942, maxDocs=44218)
              0.046056706 = queryNorm
            0.14413087 = fieldWeight in 1196, product of:
              4.0 = tf(freq=16.0), with freq of:
                16.0 = termFreq=16.0
              1.153047 = idf(docFreq=37942, maxDocs=44218)
              0.03125 = fieldNorm(doc=1196)
      0.5 = coord(1/2)
  0.5 = coord(1/2)

Abstract

This article describes the journey from the conception of and vision for a modern search-engine-based search environment to its technological realisation. In doing so, it takes up the thread of an earlier article on this subject, this time from a technical viewpoint. As well as presenting the conceptual considerations of the initial stages, this article will principally elucidate the technological aspects of this journey. The starting point for the deliberations about development of an academic search engine was the experience we gained through the generally successful project "Digital Library NRW", in which from 1998 to 2000-with Bielefeld University Library in overall charge-we designed a system model for an Internet-based library portal with an improved academic search environment at its core. At the heart of this system was a metasearch with an availability function, to which we added a user interface integrating all relevant source material for study and research. The deficiencies of this approach were felt soon after the system was launched in June 2001. There were problems with the stability and performance of the database retrieval system, with the integration of full-text documents and Internet pages, and with acceptance by users, because users are increasingly performing the searches themselves using search engines rather than going to the library for help in doing searches. Since a long list of problems are also encountered using commercial search engines for academic use (in particular the retrieval of academic information and long-term availability), the idea was born for a search engine configured specifically for academic use. We also hoped that with one single access point founded on improved search engine technology, we could access the heterogeneous academic resources of subject-based bibliographic databases, catalogues, electronic newspapers, document servers and academic web pages.

Type

a

0.0018909799 = product of:
  0.0037819599 = sum of:
    0.0037819599 = product of:
      0.0075639198 = sum of:
        0.0075639198 = weight(_text_:a in 2548) [ClassicSimilarity], result of:
          0.0075639198 = score(doc=2548,freq=10.0), product of:
            0.053105544 = queryWeight, product of:
              1.153047 = idf(docFreq=37942, maxDocs=44218)
              0.046056706 = queryNorm
            0.14243183 = fieldWeight in 2548, product of:
              3.1622777 = tf(freq=10.0), with freq of:
                10.0 = termFreq=10.0
              1.153047 = idf(docFreq=37942, maxDocs=44218)
              0.0390625 = fieldNorm(doc=2548)
      0.5 = coord(1/2)
  0.5 = coord(1/2)

Abstract

Page Rank is a topic much discussed by Search Engine Optimisation (SEO) experts. At the heart of PageRank is a mathematical formula that seems scary to look at but is actually fairly simple to understand. Despite this many people seem to get it wrong! In particular "Chris Ridings of www.searchenginesystems.net" has written a paper entitled "PageRank Explained: Everything you've always wanted to know about PageRank", pointed to by many people, that contains a fundamental mistake early on in the explanation! Unfortunately this means some of the recommendations in the paper are not quite accurate. By showing code to correctly calculate real PageRank I hope to achieve several things in this response: - Clearly explain how PageRank is calculated. - Go through every example in Chris' paper, and add some more of my own, showing the correct PageRank for each diagram. By showing the code used to calculate each diagram I've opened myself up to peer review - mostly in an effort to make sure the examples are correct, but also because the code can help explain the PageRank calculations. - Describe some principles and observations on website design based on these correctly calculated examples. Any good web designer should take the time to fully understand how PageRank really works - if you don't then your site's layout could be seriously hurting your Google listings! [Note: I have nothing in particular against Chris. If I find any other papers on the subject I'll try to comment evenly]

Type

a