Implementazione di un sistema per la classificazione di banchi di pesce da dati di acustica

La tesi è finalizzata ad una preliminare fase di sperimentazione di un algoritmo che, a partire da dati di acustica, sia in grado di classificare le specie di pesce presenti in cale mono e plurispecifiche. I dati sono stati acquisiti nella fascia costiera della Sicilia meridionale, durante alcune campagne di ricerca effettuate tra il 2002 e il 2011, dall’IAMC – CNR di Capo Granitola. Sono stati registrati i valori delle variabili ambientali e biotiche tramite metodologia acustica e della composizione dei banchi di pesci catturati tramite cale sperimentali: acciughe, sardine, suri, altre specie pelagiche e pesci demersali. La metodologia proposta per la classificazione dei segnali acustici nasce dalla fusione di logica fuzzy e teorema di Bayes, per dar luogo ad un approccio modellistico consistente in un compilatore naïve Bayes operante in ambiente fuzzy. Nella fattispecie si è proceduto alla fase di training del classificatore, mediante un learning sample di percentuali delle categorie ittiche sopra menzionate, e ai dati di alcune delle osservazioni acustiche, biotiche e abiotiche, rilevate dall’echosurvey sugli stessi banchi. La validazione del classificatore è stata effettuata sul test set, ossia sui dati che non erano stati scelti per la fase di training. Per ciascuna cala, sono stati infine tracciati dei grafici di dispersione/correlazione dei gruppi ittici e le percentuali simulate. Come misura di corrispondenza dei dati sono stati considerati i valori di regressione R2 tra le percentuali reali e quelle calcolate dal classificatore fuzzy naïve Bayes. Questi, risultando molto alti (0,9134-0,99667), validavano il risultato del classificatore che discriminava con accuratezza le ecotracce provenienti dai banchi. L’applicabilità del classificatore va comunque testata e verificata oltre i limiti imposti da un lavoro di tesi; in particolare la fase di test va riferita a specie diverse, a condizioni ambientali al contorno differenti da quelle riscontrate e all’utilizzo di learning sample meno estesi.

Vergleichende Analyse der Textsorte Audioguide im deutschen Sprachraum und Anwendungsmöglichkeiten auf die Übersetzung eines Multimedia-Guides aus dem Italienischen ins Deutsche (Analisi contrastiva dell'audioguida come tipo testuale nelle aree germanofone e possibili applicazioni alla traduzione di una guida multimediale dall'italiano al tedesco)

La parte centrale di questo elaborato consiste nella traduzione della guida multimediale della Biblioteca Malatestiana di Cesena dall’italiano al tedesco che include un'audioguida e un Ibook. Siccome non è stato possibile reperire materiale teorico sulle convenzioni testuali e la traduzione di audioguide, è stata condotta un’analisi contrastiva di sette audioguide delle aree germanofone (Germania, Austria, Svizzera e Alto Adige) con l’obiettivo di individuare le caratteristiche testuali comuni. In una fase successiva è stato effettuato un confronto tra i risultati dell’analisi contrastiva e il testo di partenza per individuare le possibilità di applicazione dei risultati. Infine, vengono commentate anche le scelte traduttive in generale, legate soprattutto a fattori sintattici e pragmatici.

Cilengitide combined with cetuximab and platinum-based chemotherapy as first-line treatment in advanced non-small-cell lung cancer (NSCLC) patients: Results of an open-label, randomized, controlled phase II study (CERTO)

Background: This multicentre, open-label, randomized, controlled phase II study evaluated cilengitide in combination with cetuximab and platinum-based chemotherapy, compared with cetuximab and chemotherapy alone, as first-line treatment of patients with advanced non-small-cell lung cancer (NSCLC). Patients and methods: Patients were randomized 1:1:1 to receive cetuximab plus platinum-based chemotherapy alone (control), or combined with cilengitide 2000 mg 1×/week i.v. (CIL-once) or 2×/week i.v. (CIL-twice). A protocol amendment limited enrolment to patients with epidermal growth factor receptor (EGFR) histoscore ≥200 and closed the CIL-twice arm for practical feasibility issues. Primary end point was progression-free survival (PFS; independent read); secondary end points included overall survival (OS), safety, and biomarker analyses. A comparison between the CIL-once and control arms is reported, both for the total cohorts, as well as for patients with EGFR histoscore ≥200. Results: There were 85 patients in the CIL-once group and 84 in the control group. The PFS (independent read) was 6.2 versus 5.0 months for CIL-once versus control [hazard ratio (HR) 0.72; P = 0.085]; for patients with EGFR histoscore ≥200, PFS was 6.8 versus 5.6 months, respectively (HR 0.57; P = 0.0446). Median OS was 13.6 for CIL-once versus 9.7 months for control (HR 0.81; P = 0.265). In patients with EGFR ≥200, OS was 13.2 versus 11.8 months, respectively (HR 0.95; P = 0.855). No major differences in adverse events between CIL-once and control were reported; nausea (59% versus 56%, respectively) and neutropenia (54% versus 46%, respectively) were the most frequent. There was no increased incidence of thromboembolic events or haemorrhage in cilengitide-treated patients. αvβ3 and αvβ5 expression was neither a predictive nor a prognostic indicator. Conclusions: The addition of cilengitide to cetuximab/chemotherapy indicated potential clinical activity, with a trend for PFS difference in the independent-read analysis. However, the observed inconsistencies across end points suggest additional investigations are required to substantiate a potential role of other integrin inhibitors in NSCLC treatment.

Hierarchical Bayesian models in ecology: Reconstructing species interaction networks from non-homogeneous species abundance data

The relationships among organisms and their surroundings can be of immense complexity. To describe and understand an ecosystem as a tangled bank, multiple ways of interaction and their effects have to be considered, such as predation, competition, mutualism and facilitation. Understanding the resulting interaction networks is a challenge in changing environments, e.g. to predict knock-on effects of invasive species and to understand how climate change impacts biodiversity. The elucidation of complex ecological systems with their interactions will benefit enormously from the development of new machine learning tools that aim to infer the structure of interaction networks from field data. In the present study, we propose a novel Bayesian regression and multiple changepoint model (BRAM) for reconstructing species interaction networks from observed species distributions. The model has been devised to allow robust inference in the presence of spatial autocorrelation and distributional heterogeneity. We have evaluated the model on simulated data that combines a trophic niche model with a stochastic population model on a 2-dimensional lattice, and we have compared the performance of our model with L1-penalized sparse regression (LASSO) and non-linear Bayesian networks with the BDe scoring scheme. In addition, we have applied our method to plant ground coverage data from the western shore of the Outer Hebrides with the objective to infer the ecological interactions. (C) 2012 Elsevier B.V. All rights reserved.

TU Graz: Course: 707.000 Web Science and Web Technology: Lecture 10: Text Mining

This class introduces basics of web mining and information retrieval including, for example, an introduction to the Vector Space Model and Text Mining. Guest Lecturer: Dr. Michael Granitzer Optional: Modeling the Internet and the Web: Probabilistic Methods and Algorithms, Pierre Baldi, Paolo Frasconi, Padhraic Smyth, Wiley, 2003 (Chapter 4, Text Analysis)

An Animating Spark: Mundane Computing and the Web of Data

Network connectivity is reaching more and more into the physical world. This is potentially transformative – allowing every object and service in the world to talk to one other—and to their users—through any networked interface; where online services are the connective tissue of the physical world and where physical objects are avatars of online services.

Analisi del comportamento a caldo dell'acciaio AISI - H11 per la stima della vita utile di matrici per estrusione

The effect of process parameters on the creep-fatigue behavior of a hot-work tool steel for aluminum extrusion die was investigated through a technological test in which the specimen geometry resembled the mandrel of a hollow extrusion die. Tests were performed on a Gleeble thermomechanical simulator by heating the specimen using joule’s effect and by applying cyclic loading up to 6.30 h or till specimen failure. Displacements during the tests at 380, 490, 540 and 580°C and under the average stresses of 400, 600 and 800 MPa were determined. In the first set of test a dwell time of 3 min was introduced during each of the tests to understand the creep behavior. The results showed that the test could indeed physically simulate the cyclic loading on the hollow die during extrusion and reveal all the mechanisms of creep-fatigue interaction. In the second set a pure fatigue laod were induced and in the third set a static creep load were induced in the specimens. Furher type of tests, finite element and microstructural analysis were presented.

Problemi di diritto e di privacy nell’uso delle nuove tecnologie di rilevamento per immagini in ambito civile.

La presente tesi, pur senza avere pretese di completezza e rimanendo in un ambito di raccolta di informazioni e dati, intende approfondire ed analizzare gli aspetti legati alla normativa vigente ed alle problematiche di diritto e privacy per il rilevamento di immagini in ambito civile, partendo dalla descrizione generale dei più comuni sistemi di rilevamento e attraverso l’analisi di casi di studio e di esempi d’applicazione.

Linked Opinions: Describing Sentiments on the Structured Web of Data

In the paper we report on the results of our experiments on the construction of the opinion ontology. Our aim is to show the benefits of publishing in the open, on the Web, the results of the opinion mining process in a structured form. On the road to achieving this, we attempt to answer the research question to what extent opinion information can be formalized in a unified way. Furthermore, as part of the evaluation, we experiment with the usage of Semantic Web technologies and show particular use cases that support our claims.