-
Kiros, R.; Salakhutdinov, R.; Zemel, R.S.: Unifying visual-semantic embeddings with multimodal neural language models (2014)
0.25
0.25179514 = product of:
0.33572686 = sum of:
0.130858 = weight(_text_:vector in 1871) [ClassicSimilarity], result of:
0.130858 = score(doc=1871,freq=2.0), product of:
0.30654848 = queryWeight, product of:
6.439392 = idf(docFreq=191, maxDocs=44218)
0.047605187 = queryNorm
0.4268754 = fieldWeight in 1871, product of:
1.4142135 = tf(freq=2.0), with freq of:
2.0 = termFreq=2.0
6.439392 = idf(docFreq=191, maxDocs=44218)
0.046875 = fieldNorm(doc=1871)
0.17187454 = weight(_text_:space in 1871) [ClassicSimilarity], result of:
0.17187454 = score(doc=1871,freq=8.0), product of:
0.24842183 = queryWeight, product of:
5.2183776 = idf(docFreq=650, maxDocs=44218)
0.047605187 = queryNorm
0.6918657 = fieldWeight in 1871, product of:
2.828427 = tf(freq=8.0), with freq of:
8.0 = termFreq=8.0
5.2183776 = idf(docFreq=650, maxDocs=44218)
0.046875 = fieldNorm(doc=1871)
0.03299433 = product of:
0.06598866 = sum of:
0.06598866 = weight(_text_:model in 1871) [ClassicSimilarity], result of:
0.06598866 = score(doc=1871,freq=4.0), product of:
0.1830527 = queryWeight, product of:
3.845226 = idf(docFreq=2569, maxDocs=44218)
0.047605187 = queryNorm
0.36048993 = fieldWeight in 1871, product of:
2.0 = tf(freq=4.0), with freq of:
4.0 = termFreq=4.0
3.845226 = idf(docFreq=2569, maxDocs=44218)
0.046875 = fieldNorm(doc=1871)
0.5 = coord(1/2)
0.75 = coord(3/4)
- Abstract
- Inspired by recent advances in multimodal learning and machine translation, we introduce an encoder-decoder pipeline that learns (a): a multimodal joint embedding space with images and text and (b): a novel language model for decoding distributed representations from our space. Our pipeline effectively unifies joint image-text embedding models with multimodal neural language models. We introduce the structure-content neural language model that disentangles the structure of a sentence to its content, conditioned on representations produced by the encoder. The encoder allows one to rank images and sentences while the decoder can generate novel descriptions from scratch. Using LSTM to encode sentences, we match the state-of-the-art performance on Flickr8K and Flickr30K without using object detections. We also set new best results when using the 19-layer Oxford convolutional network. Furthermore we show that with linear encoders, the learned embedding space captures multimodal regularities in terms of vector space arithmetic e.g. *image of a blue car* - "blue" + "red" is near images of red cars. Sample captions generated for 800 images are made available for comparison.
-
Kiela, D.; Clark, S.: Detecting compositionality of multi-word expressions using nearest neighbours in vector space models (2013)
0.20
0.20439655 = product of:
0.4087931 = sum of:
0.24674822 = weight(_text_:vector in 1161) [ClassicSimilarity], result of:
0.24674822 = score(doc=1161,freq=4.0), product of:
0.30654848 = queryWeight, product of:
6.439392 = idf(docFreq=191, maxDocs=44218)
0.047605187 = queryNorm
0.804924 = fieldWeight in 1161, product of:
2.0 = tf(freq=4.0), with freq of:
4.0 = termFreq=4.0
6.439392 = idf(docFreq=191, maxDocs=44218)
0.0625 = fieldNorm(doc=1161)
0.16204487 = weight(_text_:space in 1161) [ClassicSimilarity], result of:
0.16204487 = score(doc=1161,freq=4.0), product of:
0.24842183 = queryWeight, product of:
5.2183776 = idf(docFreq=650, maxDocs=44218)
0.047605187 = queryNorm
0.6522972 = fieldWeight in 1161, product of:
2.0 = tf(freq=4.0), with freq of:
4.0 = termFreq=4.0
5.2183776 = idf(docFreq=650, maxDocs=44218)
0.0625 = fieldNorm(doc=1161)
0.5 = coord(2/4)
- Abstract
- We present a novel unsupervised approach to detecting the compositionality of multi-word expressions. We compute the compositionality of a phrase through substituting the constituent words with their "neighbours" in a semantic vector space and averaging over the distance between the original phrase and the substituted neighbour phrases. Several methods of obtaining neighbours are presented. The results are compared to existing supervised results and achieve state-of-the-art performance on a verb-object dataset of human compositionality ratings.
-
Rehurek, R.; Sojka, P.: Software framework for topic modelling with large corpora (2010)
0.18
0.18009433 = product of:
0.24012578 = sum of:
0.130858 = weight(_text_:vector in 1058) [ClassicSimilarity], result of:
0.130858 = score(doc=1058,freq=2.0), product of:
0.30654848 = queryWeight, product of:
6.439392 = idf(docFreq=191, maxDocs=44218)
0.047605187 = queryNorm
0.4268754 = fieldWeight in 1058, product of:
1.4142135 = tf(freq=2.0), with freq of:
2.0 = termFreq=2.0
6.439392 = idf(docFreq=191, maxDocs=44218)
0.046875 = fieldNorm(doc=1058)
0.08593727 = weight(_text_:space in 1058) [ClassicSimilarity], result of:
0.08593727 = score(doc=1058,freq=2.0), product of:
0.24842183 = queryWeight, product of:
5.2183776 = idf(docFreq=650, maxDocs=44218)
0.047605187 = queryNorm
0.34593284 = fieldWeight in 1058, product of:
1.4142135 = tf(freq=2.0), with freq of:
2.0 = termFreq=2.0
5.2183776 = idf(docFreq=650, maxDocs=44218)
0.046875 = fieldNorm(doc=1058)
0.023330513 = product of:
0.046661027 = sum of:
0.046661027 = weight(_text_:model in 1058) [ClassicSimilarity], result of:
0.046661027 = score(doc=1058,freq=2.0), product of:
0.1830527 = queryWeight, product of:
3.845226 = idf(docFreq=2569, maxDocs=44218)
0.047605187 = queryNorm
0.25490487 = fieldWeight in 1058, product of:
1.4142135 = tf(freq=2.0), with freq of:
2.0 = termFreq=2.0
3.845226 = idf(docFreq=2569, maxDocs=44218)
0.046875 = fieldNorm(doc=1058)
0.5 = coord(1/2)
0.75 = coord(3/4)
- Abstract
- Large corpora are ubiquitous in today's world and memory quickly becomes the limiting factor in practical applications of the Vector Space Model (VSM). In this paper, we identify a gap in existing implementations of many of the popular algorithms, which is their scalability and ease of use. We describe a Natural Language Processing software framework which is based on the idea of document streaming, i.e. processing corpora document after document, in a memory independent fashion. Within this framework, we implement several popular algorithms for topical inference, including Latent Semantic Analysis and Latent Dirichlet Allocation, in a way that makes them completely independent of the training corpus size. Particular emphasis is placed on straightforward and intuitive framework design, so that modifications and extensions of the methods and/or their application by interested practitioners are effortless. We demonstrate the usefulness of our approach on a real-world scenario of computing document similarities within an existing digital library DML-CZ.
-
Tomassen, S.L.: Research on ontology-driven information retrieval (2006 (?))
0.14
0.13549922 = product of:
0.27099845 = sum of:
0.18506117 = weight(_text_:vector in 4328) [ClassicSimilarity], result of:
0.18506117 = score(doc=4328,freq=4.0), product of:
0.30654848 = queryWeight, product of:
6.439392 = idf(docFreq=191, maxDocs=44218)
0.047605187 = queryNorm
0.603693 = fieldWeight in 4328, product of:
2.0 = tf(freq=4.0), with freq of:
4.0 = termFreq=4.0
6.439392 = idf(docFreq=191, maxDocs=44218)
0.046875 = fieldNorm(doc=4328)
0.08593727 = weight(_text_:space in 4328) [ClassicSimilarity], result of:
0.08593727 = score(doc=4328,freq=2.0), product of:
0.24842183 = queryWeight, product of:
5.2183776 = idf(docFreq=650, maxDocs=44218)
0.047605187 = queryNorm
0.34593284 = fieldWeight in 4328, product of:
1.4142135 = tf(freq=2.0), with freq of:
2.0 = termFreq=2.0
5.2183776 = idf(docFreq=650, maxDocs=44218)
0.046875 = fieldNorm(doc=4328)
0.5 = coord(2/4)
- Abstract
- An increasing number of recent information retrieval systems make use of ontologies to help the users clarify their information needs and come up with semantic representations of documents. A particular concern here is the integration of these semantic approaches with traditional search technology. The research presented in this paper examines how ontologies can be efficiently applied to large-scale search systems for the web. We describe how these systems can be enriched with adapted ontologies to provide both an in-depth understanding of the user's needs as well as an easy integration with standard vector-space retrieval systems. The ontology concepts are adapted to the domain terminology by computing a feature vector for each concept. Later, the feature vectors are used to enrich a provided query. The whole retrieval system is under development as part of a larger Semantic Web standardization project for the Norwegian oil & gas sector.
-
Paralic, J.; Kostial, I.: Ontology-based information retrieval (2003)
0.13
0.12719904 = product of:
0.25439808 = sum of:
0.2159047 = weight(_text_:vector in 1153) [ClassicSimilarity], result of:
0.2159047 = score(doc=1153,freq=4.0), product of:
0.30654848 = queryWeight, product of:
6.439392 = idf(docFreq=191, maxDocs=44218)
0.047605187 = queryNorm
0.7043085 = fieldWeight in 1153, product of:
2.0 = tf(freq=4.0), with freq of:
4.0 = termFreq=4.0
6.439392 = idf(docFreq=191, maxDocs=44218)
0.0546875 = fieldNorm(doc=1153)
0.038493384 = product of:
0.07698677 = sum of:
0.07698677 = weight(_text_:model in 1153) [ClassicSimilarity], result of:
0.07698677 = score(doc=1153,freq=4.0), product of:
0.1830527 = queryWeight, product of:
3.845226 = idf(docFreq=2569, maxDocs=44218)
0.047605187 = queryNorm
0.4205716 = fieldWeight in 1153, product of:
2.0 = tf(freq=4.0), with freq of:
4.0 = termFreq=4.0
3.845226 = idf(docFreq=2569, maxDocs=44218)
0.0546875 = fieldNorm(doc=1153)
0.5 = coord(1/2)
0.5 = coord(2/4)
- Abstract
- In the proposed article a new, ontology-based approach to information retrieval (IR) is presented. The system is based on a domain knowledge representation schema in form of ontology. New resources registered within the system are linked to concepts from this ontology. In such a way resources may be retrieved based on the associations and not only based on partial or exact term matching as the use of vector model presumes In order to evaluate the quality of this retrieval mechanism, experiments to measure retrieval efficiency have been performed with well-known Cystic Fibrosis collection of medical scientific papers. The ontology-based retrieval mechanism has been compared with traditional full text search based on vector IR model as well as with the Latent Semantic Indexing method.
-
Zhang, J.; Mostafa, J.; Tripathy, H.: Information retrieval by semantic analysis and visualization of the concept space of D-Lib® magazine (2002)
0.12
0.124573834 = product of:
0.16609845 = sum of:
0.05452417 = weight(_text_:vector in 1211) [ClassicSimilarity], result of:
0.05452417 = score(doc=1211,freq=2.0), product of:
0.30654848 = queryWeight, product of:
6.439392 = idf(docFreq=191, maxDocs=44218)
0.047605187 = queryNorm
0.17786475 = fieldWeight in 1211, product of:
1.4142135 = tf(freq=2.0), with freq of:
2.0 = termFreq=2.0
6.439392 = idf(docFreq=191, maxDocs=44218)
0.01953125 = fieldNorm(doc=1211)
0.09473693 = weight(_text_:space in 1211) [ClassicSimilarity], result of:
0.09473693 = score(doc=1211,freq=14.0), product of:
0.24842183 = queryWeight, product of:
5.2183776 = idf(docFreq=650, maxDocs=44218)
0.047605187 = queryNorm
0.3813551 = fieldWeight in 1211, product of:
3.7416575 = tf(freq=14.0), with freq of:
14.0 = termFreq=14.0
5.2183776 = idf(docFreq=650, maxDocs=44218)
0.01953125 = fieldNorm(doc=1211)
0.01683735 = product of:
0.0336747 = sum of:
0.0336747 = weight(_text_:model in 1211) [ClassicSimilarity], result of:
0.0336747 = score(doc=1211,freq=6.0), product of:
0.1830527 = queryWeight, product of:
3.845226 = idf(docFreq=2569, maxDocs=44218)
0.047605187 = queryNorm
0.18396176 = fieldWeight in 1211, product of:
2.4494898 = tf(freq=6.0), with freq of:
6.0 = termFreq=6.0
3.845226 = idf(docFreq=2569, maxDocs=44218)
0.01953125 = fieldNorm(doc=1211)
0.5 = coord(1/2)
0.75 = coord(3/4)
- Abstract
- In this article we present a method for retrieving documents from a digital library through a visual interface based on automatically generated concepts. We used a vocabulary generation algorithm to generate a set of concepts for the digital library and a technique called the max-min distance technique to cluster them. Additionally, the concepts were visualized in a spring embedding graph layout to depict the semantic relationship among them. The resulting graph layout serves as an aid to users for retrieving documents. An online archive containing the contents of D-Lib Magazine from July 1995 to May 2002 was used to test the utility of an implemented retrieval and visualization system. We believe that the method developed and tested can be applied to many different domains to help users get a better understanding of online document collections and to minimize users' cognitive load during execution of search tasks. Over the past few years, the volume of information available through the World Wide Web has been expanding exponentially. Never has so much information been so readily available and shared among so many people. Unfortunately, the unstructured nature and huge volume of information accessible over networks have made it hard for users to sift through and find relevant information. To deal with this problem, information retrieval (IR) techniques have gained more intensive attention from both industrial and academic researchers. Numerous IR techniques have been developed to help deal with the information overload problem. These techniques concentrate on mathematical models and algorithms for retrieval. Popular IR models such as the Boolean model, the vector-space model, the probabilistic model and their variants are well established.
Nevertheless, because thesaurus use has shown to improve retrieval, for our method we integrate functions in the search interface that permit users to explore built-in search vocabularies to improve retrieval from digital libraries. Our method automatically generates the terms and their semantic relationships representing relevant topics covered in a digital library. We call these generated terms the "concepts", and the generated terms and their semantic relationships we call the "concept space". Additionally, we used a visualization technique to display the concept space and allow users to interact with this space. The automatically generated term set is considered to be more representative of subject area in a corpus than an "externally" imposed thesaurus, and our method has the potential of saving a significant amount of time and labor for those who have been manually creating thesauri as well. Information visualization is an emerging discipline and developed very quickly in the last decade. With growing volumes of documents and associated complexities, information visualization has become increasingly important. Researchers have found information visualization to be an effective way to use and understand information while minimizing a user's cognitive load. Our work was based on an algorithmic approach of concept discovery and association. Concepts are discovered using an algorithm based on an automated thesaurus generation procedure. Subsequently, similarities among terms are computed using the cosine measure, and the associations among terms are established using a method known as max-min distance clustering. The concept space is then visualized in a spring embedding graph, which roughly shows the semantic relationships among concepts in a 2-D visual representation. The semantic space of the visualization is used as a medium for users to retrieve the desired documents. In the remainder of this article, we present our algorithmic approach of concept generation and clustering, followed by description of the visualization technique and interactive interface. The paper ends with key conclusions and discussions on future work.
-
Wei, W.; Ram, S.: Utilizing sozial bookmarking tag space for Web content discovery : a social network analysis approach (2010)
0.09
0.08925943 = product of:
0.17851886 = sum of:
0.1515791 = weight(_text_:space in 1) [ClassicSimilarity], result of:
0.1515791 = score(doc=1,freq=14.0), product of:
0.24842183 = queryWeight, product of:
5.2183776 = idf(docFreq=650, maxDocs=44218)
0.047605187 = queryNorm
0.61016816 = fieldWeight in 1, product of:
3.7416575 = tf(freq=14.0), with freq of:
14.0 = termFreq=14.0
5.2183776 = idf(docFreq=650, maxDocs=44218)
0.03125 = fieldNorm(doc=1)
0.026939759 = product of:
0.053879518 = sum of:
0.053879518 = weight(_text_:model in 1) [ClassicSimilarity], result of:
0.053879518 = score(doc=1,freq=6.0), product of:
0.1830527 = queryWeight, product of:
3.845226 = idf(docFreq=2569, maxDocs=44218)
0.047605187 = queryNorm
0.29433882 = fieldWeight in 1, product of:
2.4494898 = tf(freq=6.0), with freq of:
6.0 = termFreq=6.0
3.845226 = idf(docFreq=2569, maxDocs=44218)
0.03125 = fieldNorm(doc=1)
0.5 = coord(1/2)
0.5 = coord(2/4)
- Abstract
- Social bookmarking has gained popularity since the advent of Web 2.0. Keywords known as tags are created to annotate web content, and the resulting tag space composed of the tags, the resources, and the users arises as a new platform for web content discovery. Useful and interesting web resources can be located through searching and browsing based on tags, as well as following the user-user connections formed in the social bookmarking community. However, the effectiveness of tag-based search is limited due to the lack of explicitly represented semantics in the tag space. In addition, social connections between users are underused for web content discovery because of the inadequate social functions. In this research, we propose a comprehensive framework to reorganize the flat tag space into a hierarchical faceted model. We also studied the structure and properties of various networks emerging from the tag space for the purpose of more efficient web content discovery. The major research approach used in this research is social network analysis (SNA), together with methodologies employed in design science research. The contribution of our research includes: (i) a faceted model to categorize social bookmarking tags; (ii) a relationship ontology to represent the semantics of relationships between tags; (iii) heuristics to reorganize the flat tag space into a hierarchical faceted model using analysis of tag-tag co-occurrence networks; (iv) an implemented prototype system as proof-of-concept to validate the feasibility of the reorganization approach; (v) a set of evaluations of the social functions of the current networking features of social bookmarking and a series of recommendations as to how to improve the social functions to facilitate web content discovery.
- Content
- A Dissertation Submitted to the Faculty of the COMMITTEE ON BUSINESS ADMINISTRATION In Partial Fulfillment of the Requirements For the Degree of DOCTOR OF PHILOSOPHY WITH A MAJOR IN MANAGEMENT In the Graduate College THE UNIVERSITY OF ARIZONA. Vgl.: http://hdl.handle.net/10150/195123. Vgl. auch: https://www.semanticscholar.org/paper/Utilizing-social-bookmarking-tag-space-for-web-a-Ram-Wei/da9e7e5ee771008b741af7176d3f0d67128d1dca.
-
Dhillon, P.; Singh, M.: ¬An extended ontology model for trust evaluation using advanced hybrid ontology (2023)
0.08
0.07709426 = product of:
0.15418851 = sum of:
0.130858 = weight(_text_:vector in 981) [ClassicSimilarity], result of:
0.130858 = score(doc=981,freq=2.0), product of:
0.30654848 = queryWeight, product of:
6.439392 = idf(docFreq=191, maxDocs=44218)
0.047605187 = queryNorm
0.4268754 = fieldWeight in 981, product of:
1.4142135 = tf(freq=2.0), with freq of:
2.0 = termFreq=2.0
6.439392 = idf(docFreq=191, maxDocs=44218)
0.046875 = fieldNorm(doc=981)
0.023330513 = product of:
0.046661027 = sum of:
0.046661027 = weight(_text_:model in 981) [ClassicSimilarity], result of:
0.046661027 = score(doc=981,freq=2.0), product of:
0.1830527 = queryWeight, product of:
3.845226 = idf(docFreq=2569, maxDocs=44218)
0.047605187 = queryNorm
0.25490487 = fieldWeight in 981, product of:
1.4142135 = tf(freq=2.0), with freq of:
2.0 = termFreq=2.0
3.845226 = idf(docFreq=2569, maxDocs=44218)
0.046875 = fieldNorm(doc=981)
0.5 = coord(1/2)
0.5 = coord(2/4)
- Abstract
- In the blooming area of Internet technology, the concept of Internet-of-Things (IoT) holds a distinct position that interconnects a large number of smart objects. In the context of social IoT (SIoT), the argument of trust and reliability is evaluated in the presented work. The proposed framework is divided into two blocks, namely Verification Block (VB) and Evaluation Block (EB). VB defines various ontology-based relationships computed for the objects that reflect the security and trustworthiness of an accessed service. While, EB is used for the feedback analysis and proves to be a valuable step that computes and governs the success rate of the service. Support vector machine (SVM) is applied to categorise the trust-based evaluation. The security aspect of the proposed approach is comparatively evaluated for DDoS and malware attacks in terms of success rate, trustworthiness and execution time. The proposed secure ontology-based framework provides better performance compared with existing architectures.
-
Grant, S.: Developing cognitive architecture for modelling and simulation of cognition and error in complex tasks (1995)
0.07
0.07284519 = product of:
0.14569038 = sum of:
0.11458302 = weight(_text_:space in 3288) [ClassicSimilarity], result of:
0.11458302 = score(doc=3288,freq=2.0), product of:
0.24842183 = queryWeight, product of:
5.2183776 = idf(docFreq=650, maxDocs=44218)
0.047605187 = queryNorm
0.46124378 = fieldWeight in 3288, product of:
1.4142135 = tf(freq=2.0), with freq of:
2.0 = termFreq=2.0
5.2183776 = idf(docFreq=650, maxDocs=44218)
0.0625 = fieldNorm(doc=3288)
0.031107351 = product of:
0.062214702 = sum of:
0.062214702 = weight(_text_:model in 3288) [ClassicSimilarity], result of:
0.062214702 = score(doc=3288,freq=2.0), product of:
0.1830527 = queryWeight, product of:
3.845226 = idf(docFreq=2569, maxDocs=44218)
0.047605187 = queryNorm
0.33987316 = fieldWeight in 3288, product of:
1.4142135 = tf(freq=2.0), with freq of:
2.0 = termFreq=2.0
3.845226 = idf(docFreq=2569, maxDocs=44218)
0.0625 = fieldNorm(doc=3288)
0.5 = coord(1/2)
0.5 = coord(2/4)
- Abstract
- A cognitive architecture embodies the more general structures and mechnaisms out of which could be made a model of individual cognition in certain situation. The space of models and architectures has a number of dimensions, including: dependence on domain; level of specification; and extent of coverage of different phenomena
-
Beagle, D.: Visualizing keyword distribution across multidisciplinary c-space (2003)
0.07
0.069926426 = product of:
0.13985285 = sum of:
0.065429 = weight(_text_:vector in 1202) [ClassicSimilarity], result of:
0.065429 = score(doc=1202,freq=2.0), product of:
0.30654848 = queryWeight, product of:
6.439392 = idf(docFreq=191, maxDocs=44218)
0.047605187 = queryNorm
0.2134377 = fieldWeight in 1202, product of:
1.4142135 = tf(freq=2.0), with freq of:
2.0 = termFreq=2.0
6.439392 = idf(docFreq=191, maxDocs=44218)
0.0234375 = fieldNorm(doc=1202)
0.07442386 = weight(_text_:space in 1202) [ClassicSimilarity], result of:
0.07442386 = score(doc=1202,freq=6.0), product of:
0.24842183 = queryWeight, product of:
5.2183776 = idf(docFreq=650, maxDocs=44218)
0.047605187 = queryNorm
0.29958662 = fieldWeight in 1202, product of:
2.4494898 = tf(freq=6.0), with freq of:
6.0 = termFreq=6.0
5.2183776 = idf(docFreq=650, maxDocs=44218)
0.0234375 = fieldNorm(doc=1202)
0.5 = coord(2/4)
- Abstract
- The concept of c-space is proposed as a visualization schema relating containers of content to cataloging surrogates and classification structures. Possible applications of keyword vector clusters within c-space could include improved retrieval rates through the use of captioning within visual hierarchies, tracings of semantic bleeding among subclasses, and access to buried knowledge within subject-neutral publication containers. The Scholastica Project is described as one example, following a tradition of research dating back to the 1980's. Preliminary focus group assessment indicates that this type of classification rendering may offer digital library searchers enriched entry strategies and an expanded range of re-entry vocabularies. Those of us who work in traditional libraries typically assume that our systems of classification: Library of Congress Classification (LCC) and Dewey Decimal Classification (DDC), are descriptive rather than prescriptive. In other words, LCC classes and subclasses approximate natural groupings of texts that reflect an underlying order of knowledge, rather than arbitrary categories prescribed by librarians to facilitate efficient shelving. Philosophical support for this assumption has traditionally been found in a number of places, from the archetypal tree of knowledge, to Aristotelian categories, to the concept of discursive formations proposed by Michel Foucault. Gary P. Radford has elegantly described an encounter with Foucault's discursive formations in the traditional library setting: "Just by looking at the titles on the spines, you can see how the books cluster together...You can identify those books that seem to form the heart of the discursive formation and those books that reside on the margins. Moving along the shelves, you see those books that tend to bleed over into other classifications and that straddle multiple discursive formations. You can physically and sensually experience...those points that feel like state borders or national boundaries, those points where one subject ends and another begins, or those magical places where one subject has morphed into another..."
-
Vinyals, O.; Toshev, A.; Bengio, S.; Erhan, D.: ¬A picture is worth a thousand (coherent) words : building a natural description of images (2014)
0.06
0.06078091 = product of:
0.12156182 = sum of:
0.10795235 = weight(_text_:vector in 1874) [ClassicSimilarity], result of:
0.10795235 = score(doc=1874,freq=4.0), product of:
0.30654848 = queryWeight, product of:
6.439392 = idf(docFreq=191, maxDocs=44218)
0.047605187 = queryNorm
0.35215425 = fieldWeight in 1874, product of:
2.0 = tf(freq=4.0), with freq of:
4.0 = termFreq=4.0
6.439392 = idf(docFreq=191, maxDocs=44218)
0.02734375 = fieldNorm(doc=1874)
0.013609466 = product of:
0.027218932 = sum of:
0.027218932 = weight(_text_:model in 1874) [ClassicSimilarity], result of:
0.027218932 = score(doc=1874,freq=2.0), product of:
0.1830527 = queryWeight, product of:
3.845226 = idf(docFreq=2569, maxDocs=44218)
0.047605187 = queryNorm
0.14869452 = fieldWeight in 1874, product of:
1.4142135 = tf(freq=2.0), with freq of:
2.0 = termFreq=2.0
3.845226 = idf(docFreq=2569, maxDocs=44218)
0.02734375 = fieldNorm(doc=1874)
0.5 = coord(1/2)
0.5 = coord(2/4)
- Content
- "People can summarize a complex scene in a few words without thinking twice. It's much more difficult for computers. But we've just gotten a bit closer -- we've developed a machine-learning system that can automatically produce captions (like the three above) to accurately describe images the first time it sees them. This kind of system could eventually help visually impaired people understand pictures, provide alternate text for images in parts of the world where mobile connections are slow, and make it easier for everyone to search on Google for images. Recent research has greatly improved object detection, classification, and labeling. But accurately describing a complex scene requires a deeper representation of what's going on in the scene, capturing how the various objects relate to one another and translating it all into natural-sounding language. Many efforts to construct computer-generated natural descriptions of images propose combining current state-of-the-art techniques in both computer vision and natural language processing to form a complete image description approach. But what if we instead merged recent computer vision and language models into a single jointly trained system, taking an image and directly producing a human readable sequence of words to describe it? This idea comes from recent advances in machine translation between languages, where a Recurrent Neural Network (RNN) transforms, say, a French sentence into a vector representation, and a second RNN uses that vector representation to generate a target sentence in German. Now, what if we replaced that first RNN and its input words with a deep Convolutional Neural Network (CNN) trained to classify objects in images? Normally, the CNN's last layer is used in a final Softmax among known classes of objects, assigning a probability that each object might be in the image. But if we remove that final layer, we can instead feed the CNN's rich encoding of the image into a RNN designed to produce phrases. We can then train the whole system directly on images and their captions, so it maximizes the likelihood that descriptions it produces best match the training descriptions for each image.
Our experiments with this system on several openly published datasets, including Pascal, Flickr8k, Flickr30k and SBU, show how robust the qualitative results are -- the generated sentences are quite reasonable. It also performs well in quantitative evaluations with the Bilingual Evaluation Understudy (BLEU), a metric used in machine translation to evaluate the quality of generated sentences. A picture may be worth a thousand words, but sometimes it's the words that are most useful -- so it's important we figure out ways to translate from images to words automatically and accurately. As the datasets suited to learning image descriptions grow and mature, so will the performance of end-to-end approaches like this. We look forward to continuing developments in systems that can read images and generate good natural-language descriptions. To get more details about the framework used to generate descriptions from images, as well as the model evaluation, read the full paper here." Vgl. auch: https://news.ycombinator.com/item?id=8621658.
-
Lehmann, K.: Unser Gehirn kartiert auch Beziehungen räumlich (2015)
0.06
0.06036007 = product of:
0.12072014 = sum of:
0.101278044 = weight(_text_:space in 2146) [ClassicSimilarity], result of:
0.101278044 = score(doc=2146,freq=4.0), product of:
0.24842183 = queryWeight, product of:
5.2183776 = idf(docFreq=650, maxDocs=44218)
0.047605187 = queryNorm
0.40768576 = fieldWeight in 2146, product of:
2.0 = tf(freq=4.0), with freq of:
4.0 = termFreq=4.0
5.2183776 = idf(docFreq=650, maxDocs=44218)
0.0390625 = fieldNorm(doc=2146)
0.019442094 = product of:
0.03888419 = sum of:
0.03888419 = weight(_text_:model in 2146) [ClassicSimilarity], result of:
0.03888419 = score(doc=2146,freq=2.0), product of:
0.1830527 = queryWeight, product of:
3.845226 = idf(docFreq=2569, maxDocs=44218)
0.047605187 = queryNorm
0.21242073 = fieldWeight in 2146, product of:
1.4142135 = tf(freq=2.0), with freq of:
2.0 = termFreq=2.0
3.845226 = idf(docFreq=2569, maxDocs=44218)
0.0390625 = fieldNorm(doc=2146)
0.5 = coord(1/2)
0.5 = coord(2/4)
- Footnote
- Vgl. Original unter: http://www.sciencedirect.com/science/article/pii/S0896627315005243: "Morais Tavares, R., A. Mendelsohn, Y.Grossman, C.H. Williams, M. Shapiro, Y. Trope u. D. Schiller: A Map for Social Navigation in the Human Brain" in. Neuron 87(2015) no.1, S,231-243. [Deciphering the neural mechanisms of social behavior has propelled the growth of social neuroscience. The exact computations of the social brain, however, remain elusive. Here we investigated how the human br ain tracks ongoing changes in social relationships using functional neuroimaging. Participants were lead characters in a role-playing game in which they were to find a new home and a job through interactions with virtual cartoon characters. We found that a two-dimensional geometric model of social relationships, a "social space" framed by power and affiliation, predicted hippocampal activity. Moreover, participants who reported better social skills showed stronger covariance between hippocampal activity and "movement" through "social space." The results suggest that the hippocampus is crucial for social cognition, and imply that beyond framing physical locations, the hippocampus computes a more general, inclusive, abstract, and multidimensional cognitive map consistent with its role in episodic memory.].
-
Panzer, M.: Designing identifiers for the DDC (2007)
0.05
0.054783016 = product of:
0.10956603 = sum of:
0.042968635 = weight(_text_:space in 1752) [ClassicSimilarity], result of:
0.042968635 = score(doc=1752,freq=2.0), product of:
0.24842183 = queryWeight, product of:
5.2183776 = idf(docFreq=650, maxDocs=44218)
0.047605187 = queryNorm
0.17296642 = fieldWeight in 1752, product of:
1.4142135 = tf(freq=2.0), with freq of:
2.0 = termFreq=2.0
5.2183776 = idf(docFreq=650, maxDocs=44218)
0.0234375 = fieldNorm(doc=1752)
0.066597395 = sum of:
0.023330513 = weight(_text_:model in 1752) [ClassicSimilarity], result of:
0.023330513 = score(doc=1752,freq=2.0), product of:
0.1830527 = queryWeight, product of:
3.845226 = idf(docFreq=2569, maxDocs=44218)
0.047605187 = queryNorm
0.12745243 = fieldWeight in 1752, product of:
1.4142135 = tf(freq=2.0), with freq of:
2.0 = termFreq=2.0
3.845226 = idf(docFreq=2569, maxDocs=44218)
0.0234375 = fieldNorm(doc=1752)
0.04326688 = weight(_text_:22 in 1752) [ClassicSimilarity], result of:
0.04326688 = score(doc=1752,freq=10.0), product of:
0.16670525 = queryWeight, product of:
3.5018296 = idf(docFreq=3622, maxDocs=44218)
0.047605187 = queryNorm
0.2595412 = fieldWeight in 1752, product of:
3.1622777 = tf(freq=10.0), with freq of:
10.0 = termFreq=10.0
3.5018296 = idf(docFreq=3622, maxDocs=44218)
0.0234375 = fieldNorm(doc=1752)
0.5 = coord(2/4)
- Content
- "Although the Dewey Decimal Classification is currently available on the web to subscribers as WebDewey and Abridged WebDewey in the OCLC Connexion service and in an XML version to licensees, OCLC does not provide any "web services" based on the DDC. By web services, we mean presentation of the DDC to other machines (not humans) for uses such as searching, browsing, classifying, mapping, harvesting, and alerting. In order to build web-accessible services based on the DDC, several elements have to be considered. One of these elements is the design of an appropriate Uniform Resource Identifier (URI) structure for Dewey. The design goals of mapping the entity model of the DDC into an identifier space can be summarized as follows: * Common locator for Dewey concepts and associated resources for use in web services and web applications * Use-case-driven, but not directly related to and outlasting a specific use case (persistency) * Retraceable path to a concept rather than an abstract identification, reusing a means of identification that is already present in the DDC and available in existing metadata. We have been working closely with our colleagues in the OCLC Office of Research (especially Andy Houghton as well as Eric Childress, Diane Vizine-Goetz, and Stu Weibel) on a preliminary identifier syntax. The basic identifier format we are currently exploring is: http://dewey.info/{aspect}/{object}/{locale}/{type}/{version}/{resource} where * {aspect} is the aspect associated with an {object}-the current value set of aspect contains "concept", "scheme", and "index"; additional ones are under exploration * {object} is a type of {aspect} * {locale} identifies a Dewey translation * {type} identifies a Dewey edition type and contains, at a minimum, the values "edn" for the full edition or "abr" for the abridged edition * {version} identifies a Dewey edition version * {resource} identifies a resource associated with an {object} in the context of {locale}, {type}, and {version}
Some examples of identifiers for concepts follow: <http://dewey.info/concept/338.4/en/edn/22/> This identifier is used to retrieve or identify the 338.4 concept in the English-language version of Edition 22. <http://dewey.info/concept/338.4/de/edn/22/> This identifier is used to retrieve or identify the 338.4 concept in the German-language version of Edition 22. <http://dewey.info/concept/333.7-333.9/> This identifier is used to retrieve or identify the 333.7-333.9 concept across all editions and language versions. <http://dewey.info/concept/333.7-333.9/about.skos> This identifier is used to retrieve a SKOS representation of the 333.7-333.9 concept (using the "resource" element). There are several open issues at this preliminary stage of development: Use cases: URIs need to represent the range of statements or questions that could be submitted to a Dewey web service. Therefore, it seems that some general questions have to be answered first: What information does an agent have when coming to a Dewey web service? What kind of questions will such an agent ask? Placement of the {locale} component: It is still an open question if the {locale} component should be placed after the {version} component instead (<http://dewey.info/concept/338.4/edn/22/en>) to emphasize that the most important instantiation of a Dewey class is its edition, not its language version. From a services point of view, however, it could make more sense to keep the current arrangement, because users are more likely to come to the service with a present understanding of the language version they are seeking without knowing the specifics of a certain edition in which they are trying to find topics. Identification of other Dewey entities: The goal is to create a locator that does not answer all, but a lot of questions that could be asked about the DDC. Which entities are missing but should be surfaced for services or user agents? How will those services or agents interact with them? Should some entities be rendered in a different way as presented? For example, (how) should the DDC Summaries be retrievable? Would it be necessary to make the DDC Manual accessible through this identifier structure?"
-
Tzitzikas, Y.; Spyratos, N.; Constantopoulos, P.; Analyti, A.: Extended faceted ontologies (2002)
0.05
0.054633893 = product of:
0.109267786 = sum of:
0.08593727 = weight(_text_:space in 2280) [ClassicSimilarity], result of:
0.08593727 = score(doc=2280,freq=2.0), product of:
0.24842183 = queryWeight, product of:
5.2183776 = idf(docFreq=650, maxDocs=44218)
0.047605187 = queryNorm
0.34593284 = fieldWeight in 2280, product of:
1.4142135 = tf(freq=2.0), with freq of:
2.0 = termFreq=2.0
5.2183776 = idf(docFreq=650, maxDocs=44218)
0.046875 = fieldNorm(doc=2280)
0.023330513 = product of:
0.046661027 = sum of:
0.046661027 = weight(_text_:model in 2280) [ClassicSimilarity], result of:
0.046661027 = score(doc=2280,freq=2.0), product of:
0.1830527 = queryWeight, product of:
3.845226 = idf(docFreq=2569, maxDocs=44218)
0.047605187 = queryNorm
0.25490487 = fieldWeight in 2280, product of:
1.4142135 = tf(freq=2.0), with freq of:
2.0 = termFreq=2.0
3.845226 = idf(docFreq=2569, maxDocs=44218)
0.046875 = fieldNorm(doc=2280)
0.5 = coord(1/2)
0.5 = coord(2/4)
- Abstract
- A faceted ontology consists of a set of facets, where each facet consists of a predefined set of terms structured by a subsumption relation. We propose two extensions of faceted ontologies, which allow inferring conjunctions of terms that are valid in the underlying domain. We give a model-theoretic interpretation to these extended faceted ontologies and we provide mechanisms for inferring the valid conjunctions of terms. This inference service can be exploited for preventing errors during the indexing process and for deriving navigation trees that are suitable for browsing. The proposed scheme has several advantages by comparison to the hierarchical classification schemes that are currently used, namely: conceptual clarity: it is easier to understand, compactness: it takes less space, and scalability: the update operations can be formulated easier and be performed more efficiently.
-
Palm, F.: QVIZ : Query and context based visualization of time-spatial cultural dynamics (2007)
0.05
0.052643403 = product of:
0.10528681 = sum of:
0.08593727 = weight(_text_:space in 1289) [ClassicSimilarity], result of:
0.08593727 = score(doc=1289,freq=2.0), product of:
0.24842183 = queryWeight, product of:
5.2183776 = idf(docFreq=650, maxDocs=44218)
0.047605187 = queryNorm
0.34593284 = fieldWeight in 1289, product of:
1.4142135 = tf(freq=2.0), with freq of:
2.0 = termFreq=2.0
5.2183776 = idf(docFreq=650, maxDocs=44218)
0.046875 = fieldNorm(doc=1289)
0.019349536 = product of:
0.03869907 = sum of:
0.03869907 = weight(_text_:22 in 1289) [ClassicSimilarity], result of:
0.03869907 = score(doc=1289,freq=2.0), product of:
0.16670525 = queryWeight, product of:
3.5018296 = idf(docFreq=3622, maxDocs=44218)
0.047605187 = queryNorm
0.23214069 = fieldWeight in 1289, product of:
1.4142135 = tf(freq=2.0), with freq of:
2.0 = termFreq=2.0
3.5018296 = idf(docFreq=3622, maxDocs=44218)
0.046875 = fieldNorm(doc=1289)
0.5 = coord(1/2)
0.5 = coord(2/4)
- Abstract
- QVIZ will research and create a framework for visualizing and querying archival resources by a time-space interface based on maps and emergent knowledge structures. The framework will also integrate social software, such as wikis, in order to utilize knowledge in existing and new communities of practice. QVIZ will lead to improved information sharing and knowledge creation, easier access to information in a user-adapted context and innovative ways of exploring and visualizing materials over time, between countries and other administrative units. The common European framework for sharing and accessing archival information provided by the QVIZ project will open a considerably larger commercial market based on archival materials as well as a richer understanding of European history.
- Content
- Vortrag anlässlich des Workshops: "Extending the multilingual capacity of The European Library in the EDL project Stockholm, Swedish National Library, 22-23 November 2007".
-
Maaten, L. van den: Learning a parametric embedding by preserving local structure (2009)
0.05
0.050130073 = product of:
0.20052029 = sum of:
0.20052029 = weight(_text_:space in 3883) [ClassicSimilarity], result of:
0.20052029 = score(doc=3883,freq=8.0), product of:
0.24842183 = queryWeight, product of:
5.2183776 = idf(docFreq=650, maxDocs=44218)
0.047605187 = queryNorm
0.8071766 = fieldWeight in 3883, product of:
2.828427 = tf(freq=8.0), with freq of:
8.0 = termFreq=8.0
5.2183776 = idf(docFreq=650, maxDocs=44218)
0.0546875 = fieldNorm(doc=3883)
0.25 = coord(1/4)
- 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.
-
Bradford, R.B.: Relationship discovery in large text collections using Latent Semantic Indexing (2006)
0.05
0.04696106 = product of:
0.09392212 = sum of:
0.081022434 = weight(_text_:space in 1163) [ClassicSimilarity], result of:
0.081022434 = score(doc=1163,freq=4.0), product of:
0.24842183 = queryWeight, product of:
5.2183776 = idf(docFreq=650, maxDocs=44218)
0.047605187 = queryNorm
0.3261486 = fieldWeight in 1163, product of:
2.0 = tf(freq=4.0), with freq of:
4.0 = termFreq=4.0
5.2183776 = idf(docFreq=650, maxDocs=44218)
0.03125 = fieldNorm(doc=1163)
0.012899691 = product of:
0.025799382 = sum of:
0.025799382 = weight(_text_:22 in 1163) [ClassicSimilarity], result of:
0.025799382 = score(doc=1163,freq=2.0), product of:
0.16670525 = queryWeight, product of:
3.5018296 = idf(docFreq=3622, maxDocs=44218)
0.047605187 = queryNorm
0.15476047 = fieldWeight in 1163, product of:
1.4142135 = tf(freq=2.0), with freq of:
2.0 = termFreq=2.0
3.5018296 = idf(docFreq=3622, maxDocs=44218)
0.03125 = fieldNorm(doc=1163)
0.5 = coord(1/2)
0.5 = coord(2/4)
- Abstract
- This paper addresses the problem of information discovery in large collections of text. For users, one of the key problems in working with such collections is determining where to focus their attention. In selecting documents for examination, users must be able to formulate reasonably precise queries. Queries that are too broad will greatly reduce the efficiency of information discovery efforts by overwhelming the users with peripheral information. In order to formulate efficient queries, a mechanism is needed to automatically alert users regarding potentially interesting information contained within the collection. This paper presents the results of an experiment designed to test one approach to generation of such alerts. The technique of latent semantic indexing (LSI) is used to identify relationships among entities of interest. Entity extraction software is used to pre-process the text of the collection so that the LSI space contains representation vectors for named entities in addition to those for individual terms. In the LSI space, the cosine of the angle between the representation vectors for two entities captures important information regarding the degree of association of those two entities. For appropriate choices of entities, determining the entity pairs with the highest mutual cosine values yields valuable information regarding the contents of the text collection. The test database used for the experiment consists of 150,000 news articles. The proposed approach for alert generation is tested using a counterterrorism analysis example. The approach is shown to have significant potential for aiding users in rapidly focusing on information of potential importance in large text collections. The approach also has value in identifying possible use of aliases.
- Source
- Proceedings of the Fourth Workshop on Link Analysis, Counterterrorism, and Security, SIAM Data Mining Conference, Bethesda, MD, 20-22 April, 2006. [http://www.siam.org/meetings/sdm06/workproceed/Link%20Analysis/15.pdf]
-
Campbell, D.G.; Mayhew, A.: ¬A phylogenetic approach to bibliographic families and relationships (2017)
0.05
0.045528244 = product of:
0.09105649 = sum of:
0.07161439 = weight(_text_:space in 3875) [ClassicSimilarity], result of:
0.07161439 = score(doc=3875,freq=2.0), product of:
0.24842183 = queryWeight, product of:
5.2183776 = idf(docFreq=650, maxDocs=44218)
0.047605187 = queryNorm
0.28827736 = fieldWeight in 3875, product of:
1.4142135 = tf(freq=2.0), with freq of:
2.0 = termFreq=2.0
5.2183776 = idf(docFreq=650, maxDocs=44218)
0.0390625 = fieldNorm(doc=3875)
0.019442094 = product of:
0.03888419 = sum of:
0.03888419 = weight(_text_:model in 3875) [ClassicSimilarity], result of:
0.03888419 = score(doc=3875,freq=2.0), product of:
0.1830527 = queryWeight, product of:
3.845226 = idf(docFreq=2569, maxDocs=44218)
0.047605187 = queryNorm
0.21242073 = fieldWeight in 3875, product of:
1.4142135 = tf(freq=2.0), with freq of:
2.0 = termFreq=2.0
3.845226 = idf(docFreq=2569, maxDocs=44218)
0.0390625 = fieldNorm(doc=3875)
0.5 = coord(1/2)
0.5 = coord(2/4)
- Abstract
- This presentation applies the principles of phylogenetic classification to the phenomenon of bibliographic relationships in library catalogues. We argue that while the FRBR paradigm supports hierarchical bibliographic relationships between works and their various expressions and manifestations, we need a different paradigm to support associative bibliographic relationships of the kind detected in previous research. Numerous studies have shown the existence and importance of bibliographic relationships that lie outside that hierarchical FRBR model: particularly the importance of bibliographic families. We would like to suggest phylogenetics as a potential means of gaining access to those more elusive and ephemeral relationships. Phylogenetic analysis does not follow the Platonic conception of an abstract work that gives rise to specific instantiations; rather, it tracks relationships of kinship as they evolve over time. We use two examples to suggest ways in which phylogenetic trees could be represented in future library catalogues. The novels of Jane Austen are used to indicate how phylogenetic trees can represent, with greater accuracy, the line of Jane Austen adaptations, ranging from contemporary efforts to complete her unfinished work, through to the more recent efforts to graft horror memes onto the original text. Stanley Kubrick's 2001: A Space Odyssey provides an example of charting relationships both backwards and forwards in time, across different media and genres. We suggest three possible means of applying phylogenetic s in the future: enhancement of the relationship designators in RDA, crowdsourcing user tags, and extracting relationship trees through big data analysis.
-
Cecchini, C.; Zanchetta, C.; Paolo Borin, P.; Xausa, G.: Computational design e sistemi di classificazione per la verifica predittiva delle prestazioni di sistema degli organismi edilizi : Computational design and classification systems to support predictive checking of performance of building systems (2017)
0.05
0.045528244 = product of:
0.09105649 = sum of:
0.07161439 = weight(_text_:space in 5856) [ClassicSimilarity], result of:
0.07161439 = score(doc=5856,freq=2.0), product of:
0.24842183 = queryWeight, product of:
5.2183776 = idf(docFreq=650, maxDocs=44218)
0.047605187 = queryNorm
0.28827736 = fieldWeight in 5856, product of:
1.4142135 = tf(freq=2.0), with freq of:
2.0 = termFreq=2.0
5.2183776 = idf(docFreq=650, maxDocs=44218)
0.0390625 = fieldNorm(doc=5856)
0.019442094 = product of:
0.03888419 = sum of:
0.03888419 = weight(_text_:model in 5856) [ClassicSimilarity], result of:
0.03888419 = score(doc=5856,freq=2.0), product of:
0.1830527 = queryWeight, product of:
3.845226 = idf(docFreq=2569, maxDocs=44218)
0.047605187 = queryNorm
0.21242073 = fieldWeight in 5856, product of:
1.4142135 = tf(freq=2.0), with freq of:
2.0 = termFreq=2.0
3.845226 = idf(docFreq=2569, maxDocs=44218)
0.0390625 = fieldNorm(doc=5856)
0.5 = coord(1/2)
0.5 = coord(2/4)
- Abstract
- The aim of control the economic, social and environmental aspects connected to the construction of a building imposes a systematic approach for which t is necessary to make test models aimed to a coordinate analysis of different and independent performance issues. BIM technology, referring to interoperable informative models, offers a significant operative basis to achieve this necessity. In most of the cases, informative models concentrate on a product-based digital models collection built in a virtual space, more than on the simulation of their relational behaviors. This relation, instead, is the most important aspect of modelling because it marks and characterizes the interactions that can define the building as a system. This study presents the use of standard classification systems as tools for both the activation and validation of an integrated performance-based building process. By referring categories and types of the informative model to the codes of a technological and performance-based classification system, it is possible to link and coordinate functional units and their elements with the indications required by the AEC standards. In this way, progressing with an incremental logic, it is possible to achieve the management of the requirements of the whole building and the monitoring of the fulfilment of design objectives and specific normative guidelines.
-
Yang, Y.; Liu, X.: ¬A re-examination of text categorization methods (1999)
0.04
0.038166918 = product of:
0.15266767 = sum of:
0.15266767 = weight(_text_:vector in 3386) [ClassicSimilarity], result of:
0.15266767 = score(doc=3386,freq=2.0), product of:
0.30654848 = queryWeight, product of:
6.439392 = idf(docFreq=191, maxDocs=44218)
0.047605187 = queryNorm
0.4980213 = fieldWeight in 3386, product of:
1.4142135 = tf(freq=2.0), with freq of:
2.0 = termFreq=2.0
6.439392 = idf(docFreq=191, maxDocs=44218)
0.0546875 = fieldNorm(doc=3386)
0.25 = coord(1/4)
- Abstract
- This paper reports a controlled study with statistical significance tests an five text categorization methods: the Support Vector Machines (SVM), a k-Nearest Neighbor (kNN) classifier, a neural network (NNet) approach, the Linear Leastsquares Fit (LLSF) mapping and a Naive Bayes (NB) classifier. We focus an the robustness of these methods in dealing with a skewed category distribution, and their performance as function of the training-set category frequency. Our results show that SVM, kNN and LLSF significantly outperform NNet and NB when the number of positive training instances per category are small (less than ten, and that all the methods perform comparably when the categories are sufficiently common (over 300 instances).
