Suffix Tree Clustering (STC) is a search result clustering (SRC) algorithm focused on generating overlapping clusters with meaningful labels in linear time. It showed the feasibility of SRC but in time, subsequent studies introduced description-first algorithms that generate better labels and achieve higher precision. Still, STC remained as the fastest SRC algorithm and there appeared studies concerned with different problems of STC. In this thesis, semantic relations between cluster labels and documents are exploited to filter out noisy labels and improve merging phase of STC. Wikipedia is used to identify these relations and methods for integrating semantic information to STC are suggested. Semantic features are shown to be effective for SRC task when used together with term frequency vectors. Furthermore, there were no SRC studies on Turkish up to now. In this thesis, a dataset for Turkish is introduced and a number of methods are tested on Turkish.

Due to the ever-increasing volume of textual documents, text mining is a rapidly growing application of knowledge discovery in databases. Past text mining techniques predominately concentrated on discovering intra-document patterns from textual documents, such as text categorization, document clustering, query expansion, and event tracking. Mining inter-document patterns from textual documents has been largely ignored in the literature. This research focuses on discovering inter-document patterns, called evolution patterns, from document-sequences and proposed the evolution pattern discovery (EPD) technique for mining evolution patterns from a set of ordered sequences of documents. The discovery of evolution patterns can be applied in such domains as environmental scanning and knowledge management, and can be used to facilitate existing document management and retrieval techniques (e.g., event tracking).

Recent advances in computer and network technologies have contributed significantly to global connectivity and stimulated the amount of online textual document to grow extremely rapidly. The rapid accumulation of textual documents on the Web or within an organization requires effective document management techniques, covering from information retrieval, information filtering and text mining. The word mismatch problem represents a challenging issue to be addressed by the document management research. Word mismatch has been extensively investigated in information retrieval (IR) research by the use of term expansion (or specifically query expansion). However, a review of text mining literature suggests that the word mismatch problem has seldom been addressed by text mining techniques. Thus, this thesis aims at investigating the use of term expansion on some text mining techniques, specifically including text categorization, document clustering and event detection. Accordingly, we developed term expansion extensions to these three text mining techniques. The empirical evaluation results showed that term expansion increased the categorization effectiveness when the correlation coefficient feature selection was employed. With respect to document clustering, techniques extended with term expansion achieved comparable clustering effectiveness to existing techniques and showed its superiority in improving clustering specificity measure. Finally, the use of term expansion for supporting event detection has degraded the detection effectiveness as compared to the traditional event detection technique.

A term weighting scheme measures the importance of a term in a collection. A document ranking model uses these term weights to find the rank or score of a document in a collection. We present a series of cluster-based term weighting and document ranking models based on the TF-IDF and Okapi BM25 models. These term weighting and document ranking models update the inter-cluster and intra-cluster frequency components based on the generated clusters. These inter-cluster and intra-cluster frequency components are used for weighting the importance of a term in addition to the term and document frequency components. In this thesis, we will show how these models outperform the TF-IDF and Okapi BM25 models in document clustering and ranking.

Thesis (Ph. D.)--Ohio State University, 2008. / Title from first page of PDF file. Includes bibliographical references (p. 168-184).

Τα τελευταία χρόνια, υπάρχει ένα αυξανόμενο ενδιαφέρον για την αυτόματη εξόρυξη κειμένων (Text Mining) με βιοϊατρικό περιεχόμενο, λόγω της ραγδαίας αύξησης των δημοσιεύσεων που είναι αποθηκευμένες σε ηλεκτρονική μορφή σε Βάσεις Δεδομένων του Παγκόσμιου Ιστού, όπως το PubMed και το Springerlink. Το βασικό πρόβλημα που κάνει αυτό τον στόχο περισσότερο προκλητικό και δύσκολο είναι η αδυναμία της επεξεργασίας της διαθέσιμης αυτής πληροφορίας και της εξαγωγής χρήσιμων συνδέσεων και συμπερασμάτων. Κρίνεται, επομένως, επιτακτική η ανάπτυξη νέων εργαλείων που θα διευκολύνουν την εξόρυξη γνώσης από κείμενα βιολογικού περιεχομένου. Σκοπός της παρούσας διπλωματικής εργασίας είναι αρχικά η παρουσίαση γνωστών μεθόδων εξόρυξης δεδομένων από κείμενα αλλά και η ανάπτυξη ενός εργαλείου για την αποδοτική και αξιόπιστη ανακάλυψη γνώσεων από βιοϊατρική βιβλιογραφία που να βασίζεται σε προηγμένες τεχνικές εξόρυξης γνώσης από κείμενα. Πιο συγκεκριμένα, η προσπάθειά μας επικεντρώνεται στην ανάπτυξη ενός αποδοτικού αλγόριθμου συσταδοποίησης και τη χρήση αποδοτικών τεχνικών που αξιολογούν τα αποτελέσματα της συσταδοποίησης, έτσι ώστε να παρέχεται βοήθεια στον χρήστη στην προσπάθεια αναζήτησης του για πληροφορία βιολογικού περιεχομένου. Ο προτεινόμενος αλγόριθμος βασίζεται σε διαφορετικές τεχνικές συσταδοποίησης, όπως ο Ιεραρχικός Αλγόριθμος και ο Spherical K-means Αλγόριθμος και εφαρμόζει μια τελική ταξινόμηση με βάση το Impact Factor των κειμένων που ανακτήθηκαν. Τα βασικά βήματα που περιλαμβάνει ο αλγόριθμος είναι: η προεπεξεργασία των κειμένων, η αναπαράσταση των κειμένων σε διανυσματική μορφή με χρήση του Διανυσματικού Μοντέλου (Vector Space Model), η εφαρμογή της Λανθάνουσας Σημασιολογικής Δεικτοδότησης (Latent Semantic Indexing), η Ασαφής Συσταδοποίηση (Fuzzy Clustering), ο Ιεραρχικός Αλγόριθμος (Hierarchical Algorithm), o Spherical K-means Αλγόριθμος, η επιλογή της καλύτερης συστάδας και τέλος η ταξινόμηση με βάση το Impact Factor των κειμένων που ανακτήθηκαν. Η εφαρμογή που υλοποιούμε βασίζεται στον παραπάνω αλγόριθμο και προσφέρει δύο τρόπους αναζήτησης: 1) σε τρέχοντα ερωτήματα του χρήστη, τα οποία αποθηκεύονται στη βάση δεδομένων και επομένως λειτουργεί ως μέσο συμπιεσμένης αποθήκευσης των προηγούμενων ερωτημάτων του χρήστη, 2) αναζήτηση μέσα από μία λίστα προκαθορισμένων Topic βιολογικού περιεχομένου και επομένως παρέχει στο χρήστη μια επιπλέον βοήθεια σε ένα ευρύ φάσμα ερωτημάτων. Επιπλέον, η εφαρμογή εξάγει χρήσιμες συσχετίσεις όρων χρησιμοποιώντας τις τελικές συστάδες. / There is an increasing interest in automatic text mining in biomedical texts due to the increasing number of electronically available publications stored in databases such as PubMed and SpringerLink. The main problem that makes this goal more challenging and difficult is the inability of processing the available information and extracting useful connections and assumptions. Therefore, there is an urgent need for new text-mining tools to facilitate the process of text mining from biomedical documents. The goal of the present diploma thesis is to present known methods of text mining, and to develop an application that provides reliable knowledge from biomedical literature based on efficient text mining techniques. In particular, our attempt is mainly focused on developing an efficient clustering algorithm and using techniques for evaluating the results of clustering, in order to assist the users in their biological information seeking activities. The proposed algorithm involves different clustering techniques, such as Hierarchical Algorithm, Spherical K-means Algorithm and employs a final ranking according to Impact Factor of retrieved documents. The basic steps of our algorithm are: preprocessing of text’s content, representation with the vector space model, applying Latent Semantic Indexing (LSI), fuzzy clustering, hierarchical clustering, spherical k-means clustering, selection of the best cluster and ranking of biomedical documents according to their impact factor. The application that we implement is based on the above algorithm and provides two search methods: 1) search with user’s queries, which are saved in the database and thus playing the role of a compacted storage of his past search activities, 2) search through a list of pre-specified biological Topics, and thus providing the user with an extra assistance in his various queries. Moreover the whole scheme can mine useful associations between terms by exploiting the nature of the formed clusters.

Indiana University-Purdue University Indianapolis (IUPUI) / Biomedical data exists in the form of journal articles, research studies, electronic health records, care guidelines, etc. While text mining and natural language processing tools have been widely employed across various domains, these are just taking off in the healthcare space. A primary hurdle that makes it difficult to build artificial intelligence models that use biomedical data, is the limited amount of labelled data available. Since most models rely on supervised or semi-supervised methods, generating large amounts of pre-processed labelled data that can be used for training purposes becomes extremely costly. Even for datasets that are labelled, the lack of normalization of biomedical concepts further affects the quality of results produced and limits the application to a restricted dataset. This affects reproducibility of the results and techniques across datasets, making it difficult to deploy research solutions to improve healthcare services. The research presented in this thesis focuses on reducing the need to create labels for biomedical text mining by using unsupervised recurrent neural networks. The proposed method utilizes word embeddings to generate vector representations of biomedical concepts based on semantics and context. Experiments with unsupervised clustering of these biomedical concepts show that concepts that are similar to each other are clustered together. While this clustering captures different synonyms of the same concept, it also captures the similarities between various diseases and the symptoms that those diseases are symptomatic of. To test the performance of the concept vectors on corpora of documents, a document vector generation method that utilizes these concept vectors is also proposed. The document vectors thus generated are used as an input to clustering algorithms, and the results show that across multiple corpora, the proposed methods of concept and document vector generation outperform the baselines and provide more meaningful clustering. The applications of this document clustering are huge, especially in the search and retrieval space, providing clinicians, researchers and patients more holistic and comprehensive results than relying on the exclusive term that they search for. At the end, a framework for extracting clinical information that can be mapped to electronic health records from preventive care guidelines is presented. The extracted information can be integrated with the clinical decision support system of an electronic health record. A visualization tool to better understand and observe patient trajectories is also explored. Both these methods have potential to improve the preventive care services provided to patients.

As large amounts of digital information become more and more accessible, the ability to effectively find relevant information is increasingly important. Search engines have historically performed well at finding relevant information by relying primarily on lexical and word based measures. Similarly, standard approaches to organizing and categorizing large amounts of textual information have previously relied on lexical and word based measures to perform grouping or classification tasks. Quite often, however, these processes take place without respect to semantics, or word meanings. This is perhaps due to the fact that the idea of meaningful similarity is naturally qualitative, and thus difficult to incorporate into quantitative processes. In this thesis we formally present a method for computing quantitative document-level semantic distance, which is designed to model the degree to which humans would associate two documents with respect to conceptual similarity. We show how this metric can be applied to document retrieval and clustering problems. We conclude that while our metric is not well suited for text indexing, the use of our semantic distance metric can improve document retrieval through result set re-ranking and query expansion. We also conclude that our semantic distance metric can be used to improve document clustering in distance-based clustering algorithms.

No description available.

Written notation has a long history in many musical traditions and has been particularly important in the composition and performance of Western art music. This study adopted the conceptual view that a musical score consists of two coordinated but separate communication channels: the musical text and a collection of composer-selected performance marks that serve as an interpretive gloss on that text. Structurally, these channels are defined by largely disjoint vocabularies of symbols and words. While the sound structures represented by musical texts are well studied in music theory and analysis, the stylistic patterns of performance marks and how they acquire contextual meaning in performance is an area with fewer theoretical foundations. This quantitative research explored the possibility that composers exhibit recurring patterns in their use of performance marks. Seventeen solo piano sonatas written between 1798 and 1913 by five major composers were analyzed from modern editions by tokenizing and tabulating the types and usage frequencies of their individual performance marks without regard to the associated musical texts. Using analytic methods common in information science, the results demonstrated persistent statistical similarities among the works of each composer and differences among the work groups of different composers. Although based on a small sample, the results still offered statistical support for the existence of recurring stylistic patterns in composers' use of performance marks across their works.