Mémoire numérisé par la Division de la gestion de documents et des archives de l'Université de Montréal.

[pt] Sistemas de recomendação têm sido amplamente utilizados pelos grandes portais na Web, em decorrência do aumento do volume de dados disponíveis na Web. Tais sistemas são basicamente utilizados para sugerir informações relevantes para os seus usuários. Esta dissertação apresenta um sistema de recomendação de segundo nível para auxiliar equipes de jornalistas de portais de notícias no processo de recomendação de notícias relacionadas para os usuários do portal. O sistema é chamado de segundo nível pois apresenta recomendações aos jornalistas para que, por sua vez, geram recomendações aos usuários do portal. O modelo seguido pelo sistema consiste na recomendação de notícias relacionadas com base em características extraídas do próprio texto da notícia original. As características extraídas permitem a criação de consultas contra um banco de dados de notícias anteriormente publicadas. O resultado de uma consulta é uma lista de notícias candidatas à recomendação, ordenada pela similaridade com a notícia original e pela data de publicação, que o editor da notícia original manualmente processa para gerar a lista final de notícias relacionadas. / [en] Recommendation systems are widely used by major Web portals due to the increase in the volume of data available on the Web. Such systems are basically used to suggest information relevant to their users. This dissertation presents a second-level recommendation system, which aims at assisting the team of journalists of a news Web portal in the process of recommending related news for the users of the Web portal. The system is called second level since it creates recommendations to the journalists Who, in turn, generate recommendations to the users. The system follows a model based on features extracted from the text itself. The extracted features permit creating queries against a news database. The query result is a list of candidate news, sorted by score and date of publication, which the news editor manually processes to generate the final list of related news.

Information Extraction (IE) is the process of extracting structured and important pieces of information from a set of unstructured text documents in natural language. The final goal of structured information extraction is to populate a database and reach data effectively. Our study focuses on named entity recognition (NER) which is an important subtask of IE. NER is the task that deals with extraction of named entities like person, location, organization names, temporal expressions (date and time) and numerical expressions (money and percent). NER research on Turkish is known to be rare. There are rule-based, learning based and hybrid systems for NER on Turkish texts. Some of the learning approaches used for NER in Turkish are conditional random fields (CRF), rote learning, rule extraction and generalization. In this thesis, we propose a learning based named entity recognizer for Turkish texts which employs a modified version of Bayesian learning as the learning scheme. To the best of our knowledge, this is the first learning based system that uses Bayesian approach for NER in Turkish. Several features (like token length, capitalization, lexical meaning, etc.) are used in the system to see the effects of different features on NER process. We also propose hybrid system where the Bayesian learning-based system is utilized along with a rule-based recognition system. There are two different versions of the hybrid system. Output of rule-based recognizer is utilized in different phases in these versions. We observed increase in F-Measure values for both hybrid versions. When partial scoring is active, hybrid system reached 91.44% F-Measure value / where rule-based system result is 87.43% and learning-based system result is 88.41%. The hybrid system can be improved by utilizing rule-based and learning-based components differently in the future. Hybrid system can also be improved by using different learning approaches and combining them with existing hybrid system or forming the hybrid system with a completely new approach.

Mémoire numérisé par la Division de la gestion de documents et des archives de l'Université de Montréal

The advent of the era of big data on the Web has made automatic web information extraction an essential tool in data acquisition processes. Unfortunately, automated solutions are in most cases more error prone than those created by humans, resulting in dirty and erroneous data. Automatic repair and cleaning of the extracted data is thus a necessary complement to information extraction on the Web. This thesis investigates the problem of inducing cleaning rules on web extracted data in order to (i) repair and align the data w.r.t. an original target schema, (ii) produce repairs that are as generic as possible such that different instances can benefit from them. The problem is addressed from three different angles: replace cross-site redundancy with an ensemble of entity recognisers; produce general repairs that can be encoded in the extraction process; and exploit entity-wide relations to infer common knowledge on extracted data. First, we present ROSeAnn, an unsupervised approach to integrate semantic annotators and produce a unied and consistent annotation layer on top of them. Both the diversity in vocabulary and widely varying accuracy justify the need for middleware that reconciles different annotator opinions. Considering annotators as "black-boxes" that do not require per-domain supervision allows us to recognise semantically related content in web extracted data in a scalable way. Second, we show in WADaR how annotators can be used to discover rules to repair web extracted data. We study the problem of computing joint repairs for web data extraction programs and their extracted data, providing an approximate solution that requires no per-source supervision and proves effective across a wide variety of domains and sources. The proposed solution is effective not only in repairing the extracted data, but also in encoding such repairs in the original extraction process. Third, we investigate how relationships among entities can be exploited to discover inconsistencies and additional information. We present RuDiK, a disk-based scalable solution to discover first-order logic rules over RDF knowledge bases built from web sources. We present an approach that does not limit its search space to rules that rely on "positive" relationships between entities, as in the case with traditional mining of constraints. On the contrary, it extends the search space to also discover negative rules, i.e., patterns that lead to contradictions in the data.

This thesis describes the development and in-depth empirical investigation of a method, called BootMark, for bootstrapping the marking up of named entities in textual documents. The reason for working with documents, as opposed to for instance sentences or phrases, is that the BootMark method is concerned with the creation of corpora. The claim made in the thesis is that BootMark requires a human annotator to manually annotate fewer documents in order to produce a named entity recognizer with a given performance, than would be needed if the documents forming the basis for the recognizer were randomly drawn from the same corpus. The intention is then to use the created named en- tity recognizer as a pre-tagger and thus eventually turn the manual annotation process into one in which the annotator reviews system-suggested annotations rather than creating new ones from scratch. The BootMark method consists of three phases: (1) Manual annotation of a set of documents; (2) Bootstrapping – active machine learning for the purpose of selecting which document to an- notate next; (3) The remaining unannotated documents of the original corpus are marked up using pre-tagging with revision. Five emerging issues are identified, described and empirically investigated in the thesis. Their common denominator is that they all depend on the real- ization of the named entity recognition task, and as such, require the context of a practical setting in order to be properly addressed. The emerging issues are related to: (1) the characteristics of the named entity recognition task and the base learners used in conjunction with it; (2) the constitution of the set of documents annotated by the human annotator in phase one in order to start the bootstrapping process; (3) the active selection of the documents to annotate in phase two; (4) the monitoring and termination of the active learning carried out in phase two, including a new intrinsic stopping criterion for committee-based active learning; and (5) the applicability of the named entity recognizer created during phase two as a pre-tagger in phase three. The outcomes of the empirical investigations concerning the emerging is- sues support the claim made in the thesis. The results also suggest that while the recognizer produced in phases one and two is as useful for pre-tagging as a recognizer created from randomly selected documents, the applicability of the recognizer as a pre-tagger is best investigated by conducting a user study involving real annotators working on a real named entity recognition task.

This dissertation addresses the problem of classification of entities in text represented by noun phrases. The goal of this thesis is to develop a method for automated classification of entities appearing in datasets consisting of short textual fragments. The emphasis is on unsupervised and semi-supervised methods that will allow for fine-grained character of the assigned classes and require no labeled instances for training. The set of target classes is either user-defined or determined automatically. Our initial attempt to address the entity classification problem is called Semantic Concept Mapping (SCM) algorithm. SCM maps the noun phrases representing the entities as well as the target classes to WordNet. Graph-based WordNet similarity measures are used to assign the closest class to the noun phrase. If a noun phrase does not match any WordNet concept, a Targeted Hypernym Discovery (THD) algorithm is executed. The THD algorithm extracts a hypernym from a Wikipedia article defining the noun phrase using lexico-syntactic patterns. This hypernym is then used to map the noun phrase to a WordNet synset, but it can also be perceived as the classification result by itself, resulting in an unsupervised classification system. SCM and THD algorithms were designed for English. While adaptation of these algorithms for other languages is conceivable, we decided to develop the Bag of Articles (BOA) algorithm, which is language agnostic as it is based on the statistical Rocchio classifier. Since this algorithm utilizes Wikipedia as a source of data for classification, it does not require any labeled training instances. WordNet is used in a novel way to compute term weights. It is also used as a positive term list and for lemmatization. A disambiguation algorithm utilizing global context is also proposed. We consider the BOA algorithm to be the main contribution of this dissertation. Experimental evaluation of the proposed algorithms is performed on the WordSim353 dataset, which is used for evaluation in the Word Similarity Computation (WSC) task, and on the Czech Traveler dataset, the latter being specifically designed for the purpose of our research. BOA performance on WordSim353 achieves Spearman correlation of 0.72 with human judgment, which is close to the 0.75 correlation for the ESA algorithm, to the author's knowledge the best performing algorithm for this gold-standard dataset, which does not require training data. The advantage of BOA over ESA is that it has smaller requirements on preprocessing of the Wikipedia data. While SCM underperforms on the WordSim353 dataset, it overtakes BOA on the Czech Traveler dataset, which was designed specifically for our entity classification problem. This discrepancy requires further investigation. In a standalone evaluation of THD on Czech Traveler dataset the algorithm returned a correct hypernym for 62% of entities.

This master thesis deals with the extraction of relationships between named entities in the text. In the theoretical part of the thesis, the issue of natural language representation for machine processing is discussed. Subsequently, two partial tasks of relationship extraction are defined, namely named entities recognition and classification of relationships between them, including a summary of state-of-the-art solutions. In the practical part of the thesis, system for automatic extraction of relationships between named entities from downloaded pages is designed. The classification of relationships between entities is based on the pre-trained transformers. In this thesis, four pre-trained transformers are compared, namely BERT, XLNet, RoBERTa and ALBERT.

In this master thesis are described the history and theoretical background of named-entity recognition and implementation of the system in C++ for named entity recognition and disambiguation. The system uses local disambiguation method and statistics generated from the  Wikilinks web dataset. With implemented system and with alternative implementations are performed various experiments and tests. These experiments show that the system is sufficiently successful and fast. System participates in the Entity Recognition and Disambiguation Challenge 2014.

Today's communication between operators and professional drivers takes place through direct conversations between the parties. This thesis project explores the possibility to support the operators in classifying the topic of incoming communications and which entities are affected through the use of named entity recognition and topic classifications. By developing a synthetic training dataset, a NER model and a topic classification model was developed and evaluated to achieve F1-scores of 71.4 and 61.8 respectively. These results were explained by a low variance in the synthetic dataset in comparison to a transcribed dataset from the real world which included anomalies not represented in the synthetic dataset. The aforementioned models were integrated into the dialogue framework Emora to seamlessly handle the back and forth communication and generating responses.