Catálogo de publicaciones - libros

Searching multimedia data in particular audiovisual data is still a challenging task to fulfill. The number of digital video recordings has increased dramatically as recording technology has become more affordable and network infrastructure has become easy enough to provide download and streaming solutions. But, the accessibility and traceability of its content for further use is still rather limited. In our paper we are describing and evaluating a new approach to synchronizing auxiliary text-based material as, e. g. presentation slides with lecture video recordings. Our goal is to show that the tentative transliteration is sufficient for synchronization. Different approaches to synchronize textual material with deficient transliterations of lecture recordings are discussed and evaluated in this paper. Our evaluation data-set is based on different languages and various speakers’ recordings.

Frequent disjunctive pattern is known to be a sophisticated method of text mining in a single document that satisfies , by which we can discuss efficient algorithm based on APRIORI. In this work, we propose a new by which we can extract current frequent by a weighting method for past events from a . And we discuss some experimental results.

Tree edit distance is a conventional dissimilarity measure between labeled trees. However, tree edit distance including unit-cost edit distance contains the similarity of label and that of tree structure simultaneously. Therefore, even if the label similarity between two trees that share many nodes with the same label is high, the high label similarity is hard to be recognized from their tree edit distance when their tree sizes or shapes are quite different. To overcome this flaw, we propose a novel method that obtains a label dissimilarity measure and a structural dissimilarity measure separately by decomposing unit-cost edit distance.

With vibrant and rapidly growing web, website complexity is constantly increasing, making it more difficult for users to quickly locate the information they are looking for. This, on the other hand, becomes more and more important due to the widespread reliance on the many services available on the Internet nowadays. Web mining techniques have been successfully used for quite some time, for example in search engines like Google, to facilitate retrieval of relevant information. This paper takes a different approach, as we believe that not only search engines can facilitate the task of finding the information one is looking for, but also an optimization of a website’s internal structure, which is based on previously recorded user behavior. In this paper, we will present a novel approach to identifying problematic structures in websites. This method compares user behavior, derived via web log mining techniques, to an analysis of the website’s link structure obtained by applying the Weighted PageRank algorithm (see [19]). We will then show how to use these intermediate results in order to point out problematic website structures to the website owner.

Recommender systems could be seen as an application of a data mining process in which data collection, pre-processing, building user profiles and evaluation phases are performed in order to deliver personalised recommendations. Collaborative filtering systems rely on user-to-user similarities using standard similarity measures. The symmetry of most standard similarity measures makes it difficult to differentiate users’ patterns based on their historical behaviour. That means, they are not able to distinguish between two users when one user’ behaviour is quite similar to the other but not . We have found that the -nearest neighbour algorithm may generate groups which are not necessarily homogenous. In this paper, we use an asymmetric similarity measure in order to distinguish users’ patterns. Recommendations are delivered based on the users’ historical behaviour closest to a target user. Preliminary experimental results have shown that the similarity measure used is a powerful tool for differentiating users’ patterns.

This paper describes a state-of-the-art parallel data mining solution that employs wavelet analysis for scalable outlier detection in large complex spatio-temporal data. The algorithm has been implemented on multiprocessor architecture and evaluated on real-world meteorological data. Our solution on high-performance architecture can process massive and complex spatial data at reasonable time and yields improved prediction.

This paper presents a tool for web usage mining. The aim is centered on providing a tool that facilitates the mining process rather than implement elaborated algorithms and techniques. The tool covers different phases of the CRISP-DM methodology as data preparation, data selection, modeling and evaluation. The algorithms used in the modeling phase are those implemented in the Weka project. The tool has been tested in a web site to find access and navigation patterns.

Intrusion Detection Systems (IDS’s) monitor the traffic in computer networks for detecting suspect activities. Connectionist techniques can support the development of IDS’s by modeling ‘normal’ traffic. This paper presents the application of some unsupervised neural methods to a packet dataset for the first time. This work considers three unsupervised neural methods, namely, Vector Quantization (VQ), Self-Organizing Maps (SOM) and Auto-Associative Back-Propagation (AABP) networks. The former paradigm proves quite powerful in supporting the basic space-spanning mechanism to sift normal traffic from anomalous traffic. The SOM attains quite acceptable results in dealing with some anomalies while it fails in dealing with some others. The AABP model effectively drives a nonlinear compression paradigm and eventually yields a compact visualization of the network traffic progression.

This work presents a research on the performance capabilities of an extension of the ViSOM (Visualization Induced SOM) algorithm by the use of the ensemble meta-algorithm and a later fusion process. This main fusion process has two different variants, considering two different criteria for the similarity of nodes. These criteria are Euclidean distance and similarity on Voronoi polygons. The capabilities, strengths and weakness of the different variants of the model are discussed and compared more deeply in the present work. The details of several experiments performed over different datasets applying the variants of the fusion to the ViSOM algorithm along with same variants of fusion with the SOM are included for this purpose.

To be successful with certain classification problems or knowledge discovery tasks it is not sufficient to look at the available variables at a single point in time, but their development has to be traced over a period of time. It is shown that patterns and sequences of labeled intervals represent a particularly well suited data format for this purpose. An extension of existing classifiers is proposed that enables them to handle this kind of sequential data. Compared to earlier approaches the expressiveness of the pattern language (using Allen et al.’s interval relationships) is increased, which allows the discovery of many temporal patterns common to real-world applications.