Search (43 results, page 1 of 3)

0.024373945 = product of:
  0.036560915 = sum of:
    0.00890397 = weight(_text_:a in 4482) [ClassicSimilarity], result of:
      0.00890397 = score(doc=4482,freq=10.0), product of:
        0.05209492 = queryWeight, product of:
          1.153047 = idf(docFreq=37942, maxDocs=44218)
          0.045180224 = queryNorm
        0.1709182 = fieldWeight in 4482, 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=4482)
    0.027656946 = product of:
      0.055313893 = sum of:
        0.055313893 = weight(_text_:de in 4482) [ClassicSimilarity], result of:
          0.055313893 = score(doc=4482,freq=2.0), product of:
            0.19416152 = queryWeight, product of:
              4.297489 = idf(docFreq=1634, maxDocs=44218)
              0.045180224 = queryNorm
            0.28488597 = fieldWeight in 4482, product of:
              1.4142135 = tf(freq=2.0), with freq of:
                2.0 = termFreq=2.0
              4.297489 = idf(docFreq=1634, maxDocs=44218)
              0.046875 = fieldNorm(doc=4482)
      0.5 = coord(1/2)
  0.6666667 = coord(2/3)

Abstract

This paper presents an approach for performing knowledge discovery in texts through qualitative and quantitative analyses of high-level textual characteristics. Instead of applying mining techniques on attribute values, terms or keywords extracted from texts, the discovery process works over conceptss identified in texts. Concepts represent real world events and objects, and they help the user to understand ideas, trends, thoughts, opinions and intentions present in texts. The approach combines a quasi-automatic categorisation task (for qualitative analysis) with a mining process (for quantitative analysis). The goal is to find new and useful knowledge inside a textual collection through the use of mining techniques applied over concepts (representing text content). In this paper, an application of the approach to medical records of a psychiatric hospital is presented. The approach helps physicians to extract knowledge about patients and diseases. This knowledge may be used for epidemiological studies, for training professionals and it may be also used to support physicians to diagnose and evaluate diseases.

Type

a

0.0043799505 = product of:
  0.013139851 = sum of:
    0.013139851 = weight(_text_:a in 3940) [ClassicSimilarity], result of:
      0.013139851 = score(doc=3940,freq=4.0), product of:
        0.05209492 = queryWeight, product of:
          1.153047 = idf(docFreq=37942, maxDocs=44218)
          0.045180224 = queryNorm
        0.25222903 = fieldWeight in 3940, product of:
          2.0 = tf(freq=4.0), with freq of:
            4.0 = termFreq=4.0
          1.153047 = idf(docFreq=37942, maxDocs=44218)
          0.109375 = fieldNorm(doc=3940)
  0.33333334 = coord(1/3)

Type

a

0.0040970687 = product of:
  0.012291206 = sum of:
    0.012291206 = weight(_text_:a in 3028) [ClassicSimilarity], result of:
      0.012291206 = score(doc=3028,freq=14.0), product of:
        0.05209492 = queryWeight, product of:
          1.153047 = idf(docFreq=37942, maxDocs=44218)
          0.045180224 = queryNorm
        0.23593865 = fieldWeight in 3028, product of:
          3.7416575 = tf(freq=14.0), with freq of:
            14.0 = termFreq=14.0
          1.153047 = idf(docFreq=37942, maxDocs=44218)
          0.0546875 = fieldNorm(doc=3028)
  0.33333334 = coord(1/3)

Abstract

This paper describes a system to semi-automatically extend and refine ontologies by mining textual data from the Web sites of international online media. Expanding a seed ontology creates a semantic network through co-occurrence analysis, trigger phrase analysis, and disambiguation based on the WordNet lexical dictionary. Spreading activation then processes this semantic network to find the most probable candidates for inclusion in an extended ontology. Approaches to identifying hierarchical relationships such as subsumption, head noun analysis and WordNet consultation are used to confirm and classify the found relationships. Using a seed ontology on "climate change" as an example, this paper demonstrates how spreading activation improves the result by naturally integrating the mentioned methods.

Type

a

0.0039819763 = product of:
  0.011945928 = sum of:
    0.011945928 = weight(_text_:a in 916) [ClassicSimilarity], result of:
      0.011945928 = score(doc=916,freq=18.0), product of:
        0.05209492 = queryWeight, product of:
          1.153047 = idf(docFreq=37942, maxDocs=44218)
          0.045180224 = queryNorm
        0.22931081 = fieldWeight in 916, product of:
          4.2426405 = tf(freq=18.0), with freq of:
            18.0 = termFreq=18.0
          1.153047 = idf(docFreq=37942, maxDocs=44218)
          0.046875 = fieldNorm(doc=916)
  0.33333334 = coord(1/3)

Abstract

In this paper, we present a topic discovery system aimed to reveal the implicit knowledge present in news streams. This knowledge is expressed as a hierarchy of topic/subtopics, where each topic contains the set of documents that are related to it and a summary extracted from these documents. Summaries so built are useful to browse and select topics of interest from the generated hierarchies. Our proposal consists of a new incremental hierarchical clustering algorithm, which combines both partitional and agglomerative approaches, taking the main benefits from them. Finally, a new summarization method based on Testor Theory has been proposed to build the topic summaries. Experimental results in the TDT2 collection demonstrate its usefulness and effectiveness not only as a topic detection system, but also as a classification and summarization tool.

Type

a

0.003754243 = product of:
  0.011262729 = sum of:
    0.011262729 = weight(_text_:a in 602) [ClassicSimilarity], result of:
      0.011262729 = score(doc=602,freq=16.0), product of:
        0.05209492 = queryWeight, product of:
          1.153047 = idf(docFreq=37942, maxDocs=44218)
          0.045180224 = queryNorm
        0.2161963 = fieldWeight in 602, 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=602)
  0.33333334 = coord(1/3)

Abstract

We present an approach to enhancing information access through Web structure mining in contrast to traditional approaches involving usage mining. Specifically, we mine the hardwired hierarchical hyperlink structure of Web sites to identify patterns of term-term co-occurrences we call Web functional dependencies (FDs). Intuitively, a Web FD x -> y declares that all paths through a site involving a hyperlink labeled x also contain a hyperlink labeled y. The complete set of FDs satisfied by a site help characterize (flexible and expressive) interaction paradigms supported by a site, where a paradigm is the set of explorable sequences therein. We describe algorithms for mining FDs and results from mining several hierarchical Web sites and present several interface designs that can exploit such FDs to provide compelling user experiences.

Type

a

0.0036685336 = product of:
  0.011005601 = sum of:
    0.011005601 = weight(_text_:a in 1867) [ClassicSimilarity], result of:
      0.011005601 = score(doc=1867,freq=22.0), product of:
        0.05209492 = queryWeight, product of:
          1.153047 = idf(docFreq=37942, maxDocs=44218)
          0.045180224 = queryNorm
        0.21126054 = fieldWeight in 1867, product of:
          4.690416 = tf(freq=22.0), with freq of:
            22.0 = termFreq=22.0
          1.153047 = idf(docFreq=37942, maxDocs=44218)
          0.0390625 = fieldNorm(doc=1867)
  0.33333334 = coord(1/3)

Abstract

This research creates an intelligent agent for automated collection development in a digital library setting. It uses a predictive model based an facets of each Web page to select scholarly works. The criteria came from the academic library selection literature, and a Delphi study was used to refine the list to 41 criteria. A Perl program was designed to analyze a Web page for each criterion and applied to a large collection of scholarly and nonscholarly Web pages. Bibliomining, or data mining for libraries, was then used to create different classification models. Four techniques were used: logistic regression, nonparametric discriminant analysis, classification trees, and neural networks. Accuracy and return were used to judge the effectiveness of each model an test datasets. In addition, a set of problematic pages that were difficult to classify because of their similarity to scholarly research was gathered and classified using the models. The resulting models could be used in the selection process to automatically create a digital library of Webbased scholarly research works. In addition, the technique can be extended to create a digital library of any type of structured electronic information.

Type

a

0.0035395343 = product of:
  0.010618603 = sum of:
    0.010618603 = weight(_text_:a in 4263) [ClassicSimilarity], result of:
      0.010618603 = score(doc=4263,freq=8.0), product of:
        0.05209492 = queryWeight, product of:
          1.153047 = idf(docFreq=37942, maxDocs=44218)
          0.045180224 = queryNorm
        0.20383182 = fieldWeight in 4263, 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=4263)
  0.33333334 = coord(1/3)

Abstract

Data mining (DM) is part of a process by which information can be extracted from data or databases and used to inform decision making in a variety of contexts (Benoit, 2002; Michalski, Bratka & Kubat, 1997). DM includes a range of tools and methods for extractiog information; their use in the commercial sector for knowledge extraction and discovery has been one of the main driving forces in their development (Adriaans & Zantinge, 1996; Benoit, 2002). DM has been developed and applied in numerous areas. This review describes its use in analyzing health and medical information.

Type

a

0.0035117732 = product of:
  0.010535319 = sum of:
    0.010535319 = weight(_text_:a in 2563) [ClassicSimilarity], result of:
      0.010535319 = score(doc=2563,freq=14.0), product of:
        0.05209492 = queryWeight, product of:
          1.153047 = idf(docFreq=37942, maxDocs=44218)
          0.045180224 = queryNorm
        0.20223314 = fieldWeight in 2563, product of:
          3.7416575 = tf(freq=14.0), with freq of:
            14.0 = termFreq=14.0
          1.153047 = idf(docFreq=37942, maxDocs=44218)
          0.046875 = fieldNorm(doc=2563)
  0.33333334 = coord(1/3)

Abstract

Topic discovery is an important means for marketing, e-Business and social science studies. As well, it can be applied to various purposes, such as identifying a group with certain properties and observing the emergence and diminishment of a certain cyber community. Previous topic discovery work (J.M. Kleinberg, Proceedings of the 9th Annual ACM-SIAM Symposium on Discrete Algorithms, San Francisco, California, p. 668) requires manual judgment of usefulness of outcomes and is thus incapable of handling the explosive growth of the Internet. In this paper, we propose the Automatic Topic Discovery (ATD) method, which combines a method of base set construction, a clustering algorithm and an iterative principal eigenvector computation method to discover the topics relevant to a given query without using manual examination. Given a query, ATD returns with topics associated with the query and top representative pages for each topic. Our experiments show that the ATD method performs better than the traditional eigenvector method in terms of computation time and topic discovery quality.

Type

a

0.0035117732 = product of:
  0.010535319 = sum of:
    0.010535319 = weight(_text_:a in 1330) [ClassicSimilarity], result of:
      0.010535319 = score(doc=1330,freq=14.0), product of:
        0.05209492 = queryWeight, product of:
          1.153047 = idf(docFreq=37942, maxDocs=44218)
          0.045180224 = queryNorm
        0.20223314 = fieldWeight in 1330, product of:
          3.7416575 = tf(freq=14.0), with freq of:
            14.0 = termFreq=14.0
          1.153047 = idf(docFreq=37942, maxDocs=44218)
          0.046875 = fieldNorm(doc=1330)
  0.33333334 = coord(1/3)

Abstract

This article reports on the development of a novel method for the analysis of Web logs. The method uses techniques that look for similarities between queries and identify sequences of query transformation. It allows sequences of query transformations to be represented as graphical networks, thereby giving a richer view of search behavior than is possible with the usual sequential descriptions. We also perform a basic analysis to study the correlations between observed transformation codes, with results that appear to show evidence of behavior habits. The method was developed using transaction logs from the Excite search engine to provide a tool for an ongoing research project that is endeavoring to develop a greater understanding of Web-based searching by the general public.

Type

a

0.0035117732 = product of:
  0.010535319 = sum of:
    0.010535319 = weight(_text_:a in 3144) [ClassicSimilarity], result of:
      0.010535319 = score(doc=3144,freq=14.0), product of:
        0.05209492 = queryWeight, product of:
          1.153047 = idf(docFreq=37942, maxDocs=44218)
          0.045180224 = queryNorm
        0.20223314 = fieldWeight in 3144, product of:
          3.7416575 = tf(freq=14.0), with freq of:
            14.0 = termFreq=14.0
          1.153047 = idf(docFreq=37942, maxDocs=44218)
          0.046875 = fieldNorm(doc=3144)
  0.33333334 = coord(1/3)

Abstract

Data Mining describes a technology that discovers non-trivial hidden patterns in a large collection of data. Although this technology has a tremendous impact on our lives, the invaluable contributions of this invisible technology often go unnoticed. This paper discusses advances in data mining while focusing on the emerging data mining capability. Such data mining applications perform multidimensional mining on a wide variety of heterogeneous data sources, providing solutions to many unresolved problems. This paper also highlights the advantages and disadvantages arising from the ever-expanding scope of data mining. Data Mining augments human intelligence by equipping us with a wealth of knowledge and by empowering us to perform our daily tasks better. As the mining scope and capacity increases, users and organizations become more willing to compromise privacy. The huge data stores of the 'master miners' allow them to gain deep insights into individual lifestyles and their social and behavioural patterns. Data integration and analysis capability of combining business and financial trends together with the ability to deterministically track market changes will drastically affect our lives.

Source

Social Semantic Web: Web 2.0, was nun? Hrsg.: A. Blumauer u. T. Pellegrini

Type

a

0.0034978096 = product of:
  0.010493428 = sum of:
    0.010493428 = weight(_text_:a in 2533) [ClassicSimilarity], result of:
      0.010493428 = score(doc=2533,freq=20.0), product of:
        0.05209492 = queryWeight, product of:
          1.153047 = idf(docFreq=37942, maxDocs=44218)
          0.045180224 = queryNorm
        0.20142901 = fieldWeight in 2533, product of:
          4.472136 = tf(freq=20.0), with freq of:
            20.0 = termFreq=20.0
          1.153047 = idf(docFreq=37942, maxDocs=44218)
          0.0390625 = fieldNorm(doc=2533)
  0.33333334 = coord(1/3)

Abstract

Purpose - Since library storage has been increasing day by day, it is difficult for readers to find the books which interest them as well as representative booklists. How to utilize meaningful information effectively to improve the service quality of the digital library appears to be very important. The purpose of this paper is to provide a recommendation system architecture to promote digital library services in electronic libraries. Design/methodology/approach - In the proposed architecture, a two-phase data mining process used by association rule and clustering methods is designed to generate a recommendation system. The process considers not only the relationship of a cluster of users but also the associations among the information accessed. Findings - The process considered not only the relationship of a cluster of users but also the associations among the information accessed. With the advanced filter, the recommendation supported by the proposed system architecture would be closely served to meet users' needs. Originality/value - This paper not only constructs a recommendation service for readers to search books from the web but takes the initiative in finding the most suitable books for readers as well. Furthermore, library managers are expected to purchase core and hot books from a limited budget to maintain and satisfy the requirements of readers along with promoting digital library services.

Type

a

0.003462655 = product of:
  0.010387965 = sum of:
    0.010387965 = weight(_text_:a in 3883) [ClassicSimilarity], result of:
      0.010387965 = score(doc=3883,freq=10.0), product of:
        0.05209492 = queryWeight, product of:
          1.153047 = idf(docFreq=37942, maxDocs=44218)
          0.045180224 = queryNorm
        0.19940455 = fieldWeight in 3883, 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=3883)
  0.33333334 = coord(1/3)

Abstract

The paper presents a new unsupervised dimensionality reduction technique, called parametric t-SNE, that learns a parametric mapping between the high-dimensional data space and the low-dimensional latent space. Parametric t-SNE learns the parametric mapping in such a way that the local structure of the data is preserved as well as possible in the latent space. We evaluate the performance of parametric t-SNE in experiments on three datasets, in which we compare it to the performance of two other unsupervised parametric dimensionality reduction techniques. The results of experiments illustrate the strong performance of parametric t-SNE, in particular, in learning settings in which the dimensionality of the latent space is relatively low.

Type

a

0.0033183135 = product of:
  0.0099549405 = sum of:
    0.0099549405 = weight(_text_:a in 3237) [ClassicSimilarity], result of:
      0.0099549405 = score(doc=3237,freq=18.0), product of:
        0.05209492 = queryWeight, product of:
          1.153047 = idf(docFreq=37942, maxDocs=44218)
          0.045180224 = queryNorm
        0.19109234 = fieldWeight in 3237, 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=3237)
  0.33333334 = coord(1/3)

Abstract

In this article we present a semi-supervised active learning algorithm for pattern discovery in information extraction from textual data. The patterns are reduced regular expressions composed of various characteristics of features useful in information extraction. Our major contribution is a semi-supervised learning algorithm that extracts information from a set of examples labeled as relevant or irrelevant to a given attribute. The approach is semi-supervised because it does not require precise labeling of the exact location of features in the training data. This significantly reduces the effort needed to develop a training set. An active learning algorithm is used to assist the semi-supervised learning algorithm to further reduce the training set development effort. The active learning algorithm is seeded with a Single positive example of a given attribute. The context of the seed is used to automatically identify candidates for additional positive examples of the given attribute. Candidate examples are manually pruned during the active learning phase, and our semi-supervised learning algorithm automatically discovers reduced regular expressions for each attribute. We have successfully applied this learning technique in the extraction of textual features from police incident reports, university crime reports, and patents. The performance of our algorithm compares favorably with competitive extraction systems being used in criminal justice information systems.

Type

a

0.0033183135 = product of:
  0.0099549405 = sum of:
    0.0099549405 = weight(_text_:a in 5653) [ClassicSimilarity], result of:
      0.0099549405 = score(doc=5653,freq=18.0), product of:
        0.05209492 = queryWeight, product of:
          1.153047 = idf(docFreq=37942, maxDocs=44218)
          0.045180224 = queryNorm
        0.19109234 = fieldWeight in 5653, 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=5653)
  0.33333334 = coord(1/3)

Abstract

Discusses the importance of knowledge organization in the context of the information overload caused by the vast quantities of data and information accessible on internal and external networks of an organization. Defines the characteristics of a knowledge-based product. Elaborates on the techniques and applications of text mining in developing knowledge products. Presents two approaches, as case studies, to the making of knowledge products: (1) steps and processes in the planning, designing and development of a composite multilingual multimedia CD product, with the potential international, inter-cultural end users in view, and (2) application of natural language processing software in text mining. Using a text mining software, it is possible to link concept terms from a processed text to a related thesaurus, glossary, schedules of a classification scheme, and facet structured subject representations. Concludes that the products of text mining and data mining could be made more useful if the features of a faceted scheme for subject classification are incorporated into text mining techniques and products.

Type

a

0.0033183135 = product of:
  0.0099549405 = sum of:
    0.0099549405 = weight(_text_:a in 606) [ClassicSimilarity], result of:
      0.0099549405 = score(doc=606,freq=18.0), product of:
        0.05209492 = queryWeight, product of:
          1.153047 = idf(docFreq=37942, maxDocs=44218)
          0.045180224 = queryNorm
        0.19109234 = fieldWeight in 606, 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=606)
  0.33333334 = coord(1/3)

Abstract

With more and more information available on the Internet, the task of making personalized recommendations to assist the user's navigation has become increasingly important. Considering there might be millions of users with different backgrounds accessing a Web site everyday, it is infeasible to build a separate recommendation system for each user. To address this problem, clustering techniques can first be employed to discover user groups. Then, user navigation patterns for each group can be discovered, to allow the adaptation of a Web site to the interest of each individual group. In this paper, we propose to model user access sequences as stochastic processes, and a mixture of Markov models based approach is taken to cluster users and to capture the sequential relationships inherent in user access histories. Several important issues that arise in constructing the Markov models are also addressed. The first issue lies in the complexity of the mixture of Markov models. To improve the efficiency of building/maintaining the mixture of Markov models, we develop a lightweight adapt-ive algorithm to update the model parameters without recomputing model parameters from scratch. The second issue concerns the proper selection of training data for building the mixture of Markov models. We investigate two different training data selection strategies and perform extensive experiments to compare their effectiveness on a real dataset that is generated by a Web-based knowledge management system, Livelink.

Type

a

0.0033183135 = product of:
  0.0099549405 = sum of:
    0.0099549405 = weight(_text_:a in 607) [ClassicSimilarity], result of:
      0.0099549405 = score(doc=607,freq=18.0), product of:
        0.05209492 = queryWeight, product of:
          1.153047 = idf(docFreq=37942, maxDocs=44218)
          0.045180224 = queryNorm
        0.19109234 = fieldWeight in 607, 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=607)
  0.33333334 = coord(1/3)

Abstract

Understanding what kinds of Web pages are the most useful for Web search engine users is a critical task in Web information retrieval (IR). Most previous works used hyperlink analysis algorithms to solve this problem. However, little research has been focused on query-independent Web data cleansing for Web IR. In this paper, we first provide analysis of the differences between retrieval target pages and ordinary ones based on more than 30 million Web pages obtained from both the Text Retrieval Conference (TREC) and a widely used Chinese search engine, SOGOU (www.sogou.com). We further propose a learning-based data cleansing algorithm for reducing Web pages that are unlikely to be useful for user requests. We found that there exists a large proportion of low-quality Web pages in both the English and the Chinese Web page corpus, and retrieval target pages can be identified using query-independent features and cleansing algorithms. The experimental results showed that our algorithm is effective in reducing a large portion of Web pages with a small loss in retrieval target pages. It makes it possible for Web IR tools to meet a large fraction of users' needs with only a small part of pages on the Web. These results may help Web search engines make better use of their limited storage and computation resources to improve search performance.

Type

a

0.00325127 = product of:
  0.009753809 = sum of:
    0.009753809 = weight(_text_:a in 1611) [ClassicSimilarity], result of:
      0.009753809 = score(doc=1611,freq=12.0), product of:
        0.05209492 = queryWeight, product of:
          1.153047 = idf(docFreq=37942, maxDocs=44218)
          0.045180224 = queryNorm
        0.18723148 = fieldWeight in 1611, 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=1611)
  0.33333334 = coord(1/3)

Abstract

"Garbage in, garbage out" is a well-known phrase in computer analysis, and one that comes to mind when mining Web data to draw conclusions about Web users. The challenge is that data analysts wish to infer patterns of client-side behavior from server-side data. However, because only a fraction of the user's actions ever reaches the Web server, analysts must rely an incomplete data. In this paper, we propose a client-side monitoring system that is unobtrusive and supports flexible data collection. Moreover, the proposed framework encompasses client-side applications beyond the Web browser. Expanding monitoring beyond the browser to incorporate standard office productivity tools enables analysts to derive a much richer and more accurate picture of user behavior an the Web.

Footnote

Teil eines Themenheftes: "Web retrieval and mining: A machine learning perspective"

Type

a

0.003128536 = product of:
  0.009385608 = sum of:
    0.009385608 = weight(_text_:a in 3888) [ClassicSimilarity], result of:
      0.009385608 = score(doc=3888,freq=16.0), product of:
        0.05209492 = queryWeight, product of:
          1.153047 = idf(docFreq=37942, maxDocs=44218)
          0.045180224 = queryNorm
        0.18016359 = fieldWeight in 3888, 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=3888)
  0.33333334 = coord(1/3)

Abstract

We present a new technique called "t-SNE" that visualizes high-dimensional data by giving each datapoint a location in a two or three-dimensional map. The technique is a variation of Stochastic Neighbor Embedding (Hinton and Roweis, 2002) that is much easier to optimize, and produces significantly better visualizations by reducing the tendency to crowd points together in the center of the map. t-SNE is better than existing techniques at creating a single map that reveals structure at many different scales. This is particularly important for high-dimensional data that lie on several different, but related, low-dimensional manifolds, such as images of objects from multiple classes seen from multiple viewpoints. For visualizing the structure of very large data sets, we show how t-SNE can use random walks on neighborhood graphs to allow the implicit structure of all of the data to influence the way in which a subset of the data is displayed. We illustrate the performance of t-SNE on a wide variety of data sets and compare it with many other non-parametric visualization techniques, including Sammon mapping, Isomap, and Locally Linear Embedding. The visualizations produced by t-SNE are significantly better than those produced by the other techniques on almost all of the data sets.

Type

a

0.0030970925 = product of:
  0.009291277 = sum of:
    0.009291277 = weight(_text_:a in 3835) [ClassicSimilarity], result of:
      0.009291277 = score(doc=3835,freq=2.0), product of:
        0.05209492 = queryWeight, product of:
          1.153047 = idf(docFreq=37942, maxDocs=44218)
          0.045180224 = queryNorm
        0.17835285 = fieldWeight in 3835, 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=3835)
  0.33333334 = coord(1/3)

Type

a

0.0030970925 = product of:
  0.009291277 = sum of:
    0.009291277 = weight(_text_:a in 601) [ClassicSimilarity], result of:
      0.009291277 = score(doc=601,freq=8.0), product of:
        0.05209492 = queryWeight, product of:
          1.153047 = idf(docFreq=37942, maxDocs=44218)
          0.045180224 = queryNorm
        0.17835285 = fieldWeight in 601, 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=601)
  0.33333334 = coord(1/3)

Abstract

In this paper, we present a framework for clustering Web search engine queries whose aim is to identify groups of queries used to search for similar information on the Web. The framework is based on a novel term vector model of queries that integrates user selections and the content of selected documents extracted from the logs of a search engine. The query representation obtained allows us to treat query clustering similarly to standard document clustering. We study the application of the clustering framework to two problems: relevance ranking boosting and query recommendation. Finally, we evaluate with experiments the effectiveness of our approach.

Type

a