Denna uppsats är en undersökande studie inom Information Extraction. Huvudsyftet är att skapa och utvärdera metoder inom Information Extraction och undersöka hur de kan hjälpa till att förbättra det vetenskapliga resultatet av klassificering av textelement. En deluppgift är att utvärdera den befintliga marknaden för Information Extraction i Sverige.                       För att göra detta har vi skapat ett program bestående av två delar. Den första delen utgörs av ett basfall som är en enkel metod och den andra är mer avancerad och använder sig av olika tekniker inom området Information Extraction. Fältet vi undersöker är hur ofta män och kvinnor nämns i sju olika nyhetskällor i Sverige. Resultatet jämför dessa två metoder och utvärderar dem med vetenskapliga prestationsmått inom Information Extraction.                       Studiens resultat visar på liknande förekomster av män och kvinnor mellan basfallet och den mer avancerade metoden. Undantaget är att den mer avancerade metoden har ett högre vetenskapligt värde. Marknaden för Information Extraction i Sverige är dominerad av stora medieägda bolag, där media dessutom förser dessa företag med data att analysera. Detta gör att det blir svårt att konkurrera utan en ny innovativ idé. / This paper is an investigatory report about Information Extraction. The main purpose is to create and evaluate methods within Information Extraction and see how they can help improve the scientific result in classification of text elements. A subtask is to evaluate the existing market for Information Extraction in Sweden.                       For this task a two-part computer program has been created. The first part is just a baseline with a simple method and the second one is more advanced with tools used in the field Information Extraction. The field we investigate is how often men and women are mentioned in seven different newspapers in Sweden. The result compares these two methods and evaluates them using scientific measurements of information retrieval performance.                       The results of the study show similar occurrences of men and women between the baseline and the more advanced method. The exception being that the more advanced method has a higher scientific value. The market for Information Extraction in Sweden is dominated by large corporations owned by the media, which also provide the data for these kinds of companies to analyze. This makes it hard to compete without having a new innovative idea.

Since the proposal of hypertext by Tim Berners-Lee to his employer CERN on March 12, 1989 the World Wide Web has grown to more than one billion Web pages and still grows. With the later proposed Semantic Web vision,Berners-Lee et al. suggested an extension of the existing (Document) Web to allow better reuse, sharing and understanding of data. Both the Document Web and the Web of Data (which is the current implementation of the Semantic Web) grow continuously. This is a mixed blessing, as the two forms of the Web grow concurrently and most commonly contain different pieces of information. Modern information systems must thus bridge a Semantic Gap to allow a holistic and unified access to information about a particular information independent of the representation of the data. One way to bridge the gap between the two forms of the Web is the extraction of structured data, i.e., RDF, from the growing amount of unstructured and semi-structured information (e.g., tables and XML) on the Document Web. Note, that unstructured data stands for any type of textual information like news, blogs or tweets. While extracting structured data from unstructured data allows the development of powerful information system, it requires high-quality and scalable knowledge extraction frameworks to lead to useful results. The dire need for such approaches has led to the development of a multitude of annotation frameworks and tools. However, most of these approaches are not evaluated on the same datasets or using the same measures. The resulting Evaluation Gap needs to be tackled by a concise evaluation framework to foster fine-grained and uniform evaluations of annotation tools and frameworks over any knowledge bases. Moreover, with the constant growth of data and the ongoing decentralization of knowledge, intuitive ways for non-experts to access the generated data are required. Humans adapted their search behavior to current Web data by access paradigms such as keyword search so as to retrieve high-quality results. Hence, most Web users only expect Web documents in return. However, humans think and most commonly express their information needs in their natural language rather than using keyword phrases. Answering complex information needs often requires the combination of knowledge from various, differently structured data sources. Thus, we observe an Information Gap between natural-language questions and current keyword-based search paradigms, which in addition do not make use of the available structured and unstructured data sources. Question Answering (QA) systems provide an easy and efficient way to bridge this gap by allowing to query data via natural language, thus reducing (1) a possible loss of precision and (2) potential loss of time while reformulating the search intention to transform it into a machine-readable way. Furthermore, QA systems enable answering natural language queries with concise results instead of links to verbose Web documents. Additionally, they allow as well as encourage the access to and the combination of knowledge from heterogeneous knowledge bases (KBs) within one answer. Consequently, three main research gaps are considered and addressed in this work: First, addressing the Semantic Gap between the unstructured Document Web and the Semantic Gap requires the development of scalable and accurate approaches for the extraction of structured data in RDF. This research challenge is addressed by several approaches within this thesis. This thesis presents CETUS, an approach for recognizing entity types to populate RDF KBs. Furthermore, our knowledge base-agnostic disambiguation framework AGDISTIS can efficiently detect the correct URIs for a given set of named entities. Additionally, we introduce REX, a Web-scale framework for RDF extraction from semi-structured (i.e., templated) websites which makes use of the semantics of the reference knowledge based to check the extracted data. The ongoing research on closing the Semantic Gap has already yielded a large number of annotation tools and frameworks. However, these approaches are currently still hard to compare since the published evaluation results are calculated on diverse datasets and evaluated based on different measures. On the other hand, the issue of comparability of results is not to be regarded as being intrinsic to the annotation task. Indeed, it is now well established that scientists spend between 60% and 80% of their time preparing data for experiments. Data preparation being such a tedious problem in the annotation domain is mostly due to the different formats of the gold standards as well as the different data representations across reference datasets. We tackle the resulting Evaluation Gap in two ways: First, we introduce a collection of three novel datasets, dubbed N3, to leverage the possibility of optimizing NER and NED algorithms via Linked Data and to ensure a maximal interoperability to overcome the need for corpus-specific parsers. Second, we present GERBIL, an evaluation framework for semantic entity annotation. The rationale behind our framework is to provide developers, end users and researchers with easy-to-use interfaces that allow for the agile, fine-grained and uniform evaluation of annotation tools and frameworks on multiple datasets. The decentral architecture behind the Web has led to pieces of information being distributed across data sources with varying structure. Moreover, the increasing the demand for natural-language interfaces as depicted by current mobile applications requires systems to deeply understand the underlying user information need. In conclusion, the natural language interface for asking questions requires a hybrid approach to data usage, i.e., simultaneously performing a search on full-texts and semantic knowledge bases. To close the Information Gap, this thesis presents HAWK, a novel entity search approach developed for hybrid QA based on combining structured RDF and unstructured full-text data sources.

Fakturor är en viktig källa av information för företag. Två exempel på viktiga fält i en faktura kan vara, hur mycket pengar som ska betalas och faktura id. På grund av olika format och innehåll i fakturor som skiljer sig åt är extraktionen av information från dessa fakturor ofta en manuell process som kräver mycket tid. För att kunna spara viktig information från semi-strukturerade dokument som fakturor så måste vissa företag lägga ner mycket manuellt arbete. Detta arbete inkluderar att behöva förstå fakturan och därefter veta vilket innehåll som är av intresse för företaget. Detta arbete kan ta mycket tid och därför hade en automatisering av denna process varit av stort intresse. I denna forskningen används named entity recognition för att lösa problemet. De frågor som forskningen besvarar är: Hur effektiv named entity recognition är för klassificering av rubriker i fakturor, samt hur mycket effektiviteten kan öka vid komplettering av ytterligare komponenter. Named entity recognition används för att kategorisera entiteter som i detta fallet är rubriker för fält i fakturor. Modellen som skapas ska avgöra om rubriker i fakturan kan kategoriseras under någon av kategorierna: Invoice number, invoice date, due date, customer number, total amount, vat code, vat amount eller currency. Forskningen försöker endast göra en proof of concept för att se om denna algoritm kan användas för att minska tiden av manuellt arbete. Produktionsmodellen som skapas evalueras med måttet f1-score. Den får med denna metod resultatet 79 av 100. Detta resultatet antyder på att named entity recognition kan användas i ett verkligt scenario för att identifiera rubriker av intresse i en faktura. Men för att få så bra resultat som möjligt så bör modellen kombineras med en lösning som identifierar fält med hjälp av dess data. / Invoices are an important source of information for businesses. Two examples of important fields in an invoice could be the amount of money to be paid and the invoice Id. Due to the different formats and content of invoices, the extraction of information from these is often a manual and time consuming process. In order to save important information from semi-structured documents such as invoices, some companies have to put in a lot of manual work. This work includes understanding the invoice and then knowing what content is of interest to the company. This work can take a lot of time and therefore an automation of this process would be of great interest. In this research named entity recognition is used to solve the mentioned problem. The topics for this research are: How effective named entity recognition is for classification of headers in invoices, as well as how much the efficiency can be improved by complementing with further components. Named entity recognition is used to categorize entities. In this case the entities are the headings of the invoice. The model that is created must determine whether headings in the invoice can be categorized under one of the following categories: Invoice number, invoice date, due date, customer number, total amount, vat code, vat amount or currency. This research tries to make a proof of concept to discover if this algorithm can be used to reduce the time spent on manual work. The production model that is created is evaluated with the f1-score measurement. With this method, it gets a result of 79 out of 100. This result indicates that named entity recognition can be used by companies in real-world scenarios to identify headings in invoices. But to get the best results possible, the model should also be combined with a solution that identifies fields using its corresponding data.

Hantering av ostrukturerade och fysiska dokument inom vissa områden, såsom finansiell rapportering, medför betydande ineffektivitet i dagsläget. Detta examensarbete fokuserar på utmaningen att extrahera data från ostrukturerade finansiella dokument, specifikt balansräkningar i svenska årsredovisningar, genom att använda en AI-driven pipeline. Syftet är att utveckla en metod för att automatisera datautvinning och möjliggöra förbättrad dataanalys. Projektet fokuserade på att automatisera utvinning av finansiella poster från balansräkningar genom en kombination av Optical Character Recognition (OCR) och en modell för Named Entity Recognition (NER). TesseractOCR användes för att konvertera skannade dokument till digital text, medan en BERT-baserad NER-modell tränades för att identifiera och klassificera relevanta finansiella poster. Ett Python-skript användes för att extrahera de numeriska värdena som är associerade med dessa poster. Projektet fann att NER-modellen uppnådde hög prestanda, med ett F1-score på 0,95, vilket visar dess effektivitet i att identifiera finansiella poster. Den fullständiga pipelinen lyckades extrahera över 99% av posterna från balansräkningar med en träffsäkerhet på cirka 90% för numerisk data. Projektet drar slutsatsen att kombinationen av OCR och NER är en lovande lösning för att automatisera datautvinning från ostrukturerade dokument med liknande attribut som årsredovisningar. Framtida arbeten kan utforska att förbättra träffsäkerheten i OCR och utvidga utvinningen till andra sektioner av olika typer av ostrukturerade dokument. / The persistence of unstructured and physical document management in fields such as financial reporting presents notable inefficiencies. This thesis addresses the challenge of extracting valuable data from unstructured financial documents, specifically balance sheets in Swedish annual reports, using an AI-driven pipeline. The objective is to develop a method to automate data extraction, enabling enhanced data analysis capabilities. The project focused on automating the extraction of financial posts from balance sheets using a combination of Optical Character Recognition (OCR) and a Named Entity Recognition (NER) model. TesseractOCR was used to convert scanned documents into digital text, while a fine-tuned BERT-based NER model was trained to identify and classify relevant financial features. A Python script was employed to extract the numerical values associated with these features. The study found that the NER model achieved high performance metrics, with an F1-score of 0.95, demonstrating its effectiveness in identifying financial entities. The full pipeline successfully extracted over 99% of features from balance sheets with an accuracy of about 90% for numerical data. The project concludes that combining OCR and NER technologies could be a promising solution for automating data extraction from unstructured documents with similar attributes to annual reports. Future work could explore enhancing OCR accuracy and extending the methodology to other sections of different types of unstructured documents.

Master of Science / Department of Computing and Information Sciences / William H. Hsu / Global epidemic surveillance is an essential task for national biosecurity management and bioterrorism prevention. The main goal is to protect the public from major health threads. To perform this task effectively one requires reliable, timely and accurate medical information from a wide range of sources. Towards this goal, we present a framework for epidemiological analytics that can be used to extract and visualize infectious disease outbreaks from the variety of unstructured web sources automatically. More precisely, in this thesis, we consider several research tasks including document relevance classification, entity extraction and animal disease-related event recognition in the veterinary epidemiology domain. First, we crawl web sources and classify collected documents by topical relevance using supervised learning algorithms. Next, we propose a novel approach for automated ontology construction in the veterinary medicine domain. Our approach is based on semantic relationship discovery using syntactic patterns. We then apply our automatically-constructed ontology for the domain-specific entity extraction task. Moreover, we compare our ontology-based entity extraction results with an alternative sequence labeling approach. We introduce a sequence labeling method for the entity tagging that relies on syntactic feature extraction using a sliding window. Finally, we present our novel sentence-based event recognition approach that includes three main steps: entity extraction of animal diseases, species, locations, dates and the confirmation status n-grams; event-related sentence classification into two categories - suspected or confirmed; automated event tuple generation and aggregation. We show that our document relevance classification results as well as entity extraction and disease-related event recognition results are significantly better compared to the results reported by other animal disease surveillance systems.

Since their recent introduction, conditional random fields (CRFs) have been successfully applied to a multitude of structured labelling tasks in many different domains. Examples include natural language processing (NLP), bioinformatics and computer vision. Within NLP itself we have seen many different application areas, like named entity recognition, shallow parsing, information extraction from research papers and language modelling. Most of this work has demonstrated the need, directly or indirectly, to employ some form of regularisation when applying CRFs in order to overcome the tendency for these models to overfit. To date a popular method for regularising CRFs has been to fit a Gaussian prior distribution over the model parameters. In this thesis we explore other methods of CRF regularisation, investigating their properties and comparing their effectiveness. We apply our ideas to sequence labelling problems in NLP, specifically part-of-speech tagging and named entity recognition. We start with an analysis of conventional approaches to CRF regularisation, and investigate possible extensions to such approaches. In particular, we consider choices of prior distribution other than the Gaussian, including the Laplacian and Hyperbolic; we look at the effect of regularising different features separately, to differing degrees, and explore how we may define an appropriate level of regularisation for each feature; we investigate the effect of allowing the mean of a prior distribution to take on non-zero values; and we look at the impact of relaxing the feature expectation constraints satisfied by a standard CRF, leading to a modified CRF model we call the inequality CRF. Our analysis leads to the general conclusion that although there is some capacity for improvement of conventional regularisation through modification and extension, this is quite limited. Conventional regularisation with a prior is in general hampered by the need to fit a hyperparameter or set of hyperparameters, which can be an expensive process. We then approach the CRF overfitting problem from a different perspective. Specifically, we introduce a form of CRF ensemble called a logarithmic opinion pool (LOP), where CRF distributions are combined under a weighted product. We show how a LOP has theoretical properties which provide a framework for designing new overfitting reduction schemes in terms of diverse models, and demonstrate how such diverse models may be constructed in a number of different ways. Specifically, we show that by constructing CRF models from manually crafted partitions of a feature set and combining them with equal weight under a LOP, we may obtain an ensemble that significantly outperforms a standard CRF trained on the entire feature set, and is competitive in performance to a standard CRF regularised with a Gaussian prior. The great advantage of LOP approach is that, unlike the Gaussian prior method, it does not require us to search a hyperparameter space. Having demonstrated the success of LOPs in the simple case, we then move on to consider more complex uses of the framework. In particular, we investigate whether it is possible to further improve the LOP ensemble by allowing parameters in different models to interact during training in such a way that diversity between the models is encouraged. Lastly, we show how the LOP approach may be used as a remedy for a problem that standard CRFs can sometimes suffer. In certain situations, negative effects may be introduced to a CRF by the inclusion of highly discriminative features. An example of this is provided by gazetteer features, which encode a word's presence in a gazetteer. We show how LOPs may be used to reduce these negative effects, and so provide some insight into how gazetteer features may be more effectively handled in CRFs, and log-linear models in general.

Atualmente, é comum que usuários levem em consideração a localização geográfica dos documentos, é dizer considerar o escopo geográfico que está sendo tratado no contexto do documento, nos processos de Recuperação de Informação. No entanto, os sistemas convencionais de extração de informação que estão baseados em palavras-chave não consideram que as palavras podem representar entidades geográficas espacialmente relacionadas com outras entidades nos documentos. Para resolver esse problema, é necessário viabilizar o georreferenciamento dos textos, ou seja, identificar as entidades geográficas presentes e associá-las com sua correta localização espacial. A identificação e desambiguação das entidades geográficas apresenta desafios importantes, principalmente do ponto de vista linguístico, já que um topônimo, pode possuir variados tipos de ambiguidade associados. Esse problema de ambiguidade causa ruido nos processos de recuperação de informação, já que o mesmo termo pode ter informação relevante ou irrelevante associada. Assim, a principal estratégia para superar os problemas de ambiguidade, compreende a identificação de evidências que auxiliem na identificação e desambiguação das localidades nos textos. O presente trabalho propõe uma metodologia que permite identificar e determinar a cobertura espacial dos documentos, denominada SpatialCIM. A metodologia SpatialCIM tem o objetivo de organizar os processos de resolução de topônimos. Assim, o principal objetivo deste trabalho é avaliar e selecionar técnicas de desambiguação que permitam resolver a ambiguidade dos topônimos nos textos. Para isso, foram propostas e desenvolvidas as abordagens de (1)Desambiguação por Pontos e a (2)Desambiguação Textual e Estrutural. Essas abordagens, exploram duas técnicas diferentes de desambiguação de topônimos, as quais, geram e desambiguam os caminhos geográficos associados aos topônimos reconhecidos para cada documento. Assim, a hipótese desta pesquisa é que o uso das técnicas de desambiguação de topônimos viabilizam uma melhor localização espacial dos documentos. A partir dos resultados obtidos neste trabalho, foi possível demonstrar que as técnicas de desambiguação melhoram a precisão e revocação na classificação espacial dos documentos. Demonstrou-se também o impacto positivo do uso de uma ferramenta linguística no processo de reconhecimento das entidades geográficas. Assim, foi demostrada a utilidade dos processos de desambiguação para a obtenção da cobertura espacial dos documentos / Currently, it is usual that users take into account the geographical localization of the documents in the Information Retrieval process. However, the conventional information retrieval systems based on key-word matching do not consider which words can represent geographical entities that are spatially related to other entities in the documents. To solve this problem, it is necessary to enable the geo-referencing of texts by identifying the geographical entities present in text and associate them with their correct spatial location. The identification and disambiguation of the geographical entities present major challenges mainly from the linguistic point of view, since one location can have different types of associated ambiguity. The ambiguity problem causes noise in the process of information retrieval, since the same term may have relevant or irrelevant information associated. Thus, the main strategy to overcome these problems, include the identification of evidence to assist in the identification and disambiguation of locations in the texts. This study proposes a methodology that allows the identification and spatial localization of the documents, denominated SpatialCIM. The SpatialCIM methodology has the objective to organize the Topônym Resolution process. Therefore the main objective of this study is to evaluate and select disambiguation techniques that allow solving the toponym ambiguity in texts. Therefore, we proposed and developed the approaches of (1) Disambiguation for Points and (2) Textual and Structural Disambiguation. These approaches exploit two different techniques of toponym disambiguation, which generate and desambiguate the associated paths with the recognized geographical toponym for each document. Therefore the hypothesis is, that the use of the toponyms disambiguation techniques enable a better spatial localization of documents. From the results it was possible to demonstrate that the disambiguation techniques improve the precision and recall for the spatial classification of documents. The positive effect of using a linguistic tool for the process of geographical entities recognition was also demonstrated. Thus, it was proved the usefulness of the disambiguation process for obtaining a spatial coverage of the document

In text analysis, authorship attribution occurs in a variety of ways. The field of computational linguistics becomes more important as the need of authorship attribution and text analysis becomes more widespread. For this research, pre-existing authorship attribution software, Java Graphical Authorship Attribution Program (JGAAP), implements a named entity recognizer, specifically the Stanford Named Entity Recognizer, to probe into similar genre text and to aid in extricating the correct author. This research specifically examines the words authors use around named entities in order to test the ability of these words at attributing authorship / McAnulty College and Graduate School of Liberal Arts; / Computational Mathematics / MS; / Thesis;

The information overload problem becomes imperative with the explosion of information, and people need some agents to facilitate them to filter the information to meet their personal need. In this work, we conduct a research for the article classification in the tourism domain so as to identify articles that meet users¡¦ information need. We propose an information need orientation model in tourism, which consists of four goals: Initiation, Attraction, Accommodation, and Route planning. These goals can be characterized by 13 features. Some of the identified features can be enhanced by WordNet and Named Entity Recognition techniques as supplement techniques. To test the effectiveness of using the 13 features for classification and the relevant methods, we collected 15,797 articles from TripAdvisor.com, the world's largest travel site, and randomly selected 600 articles as training data labeled by two labelers. The experimental results show that our approach generally has comparable or better performance than that of using purely lexical features, namely TF-IDF, for classification, with fewer features.

Large digital libraries typically contain large collections of heterogeneous resources intended to be delivered to a variety of user communities. One key challenge for these libraries is providing tight integration between resources both within a single collection and across the several collections of the library with out requiring hand coding. One key tool in doing this is elucidating the internal structure of the digital resources and using that structure to form connections between the resources. The heterogeneous nature of the collections and the diversity of the needs in the user communities complicates this task. Accordingly, in this thesis, I describe an approach to implementing a feature identification system to support digital collections that provides a general framework for applications while allowing decisions about the details of document representation and features identification to be deferred to domain specific implementations of that framework. These deferred decisions include details of the semantics and syntax of markup, the types of metadata to be attached to documents, the types of features to be identified, the feature identification algorithms to be applied, and which features should be indexed. This approach results in strong support for the general aspects of developing a feature identification system allowing future work to focus on the details of applying that system to the specific needs of individual collections and user communities.