Dimitrios Kokkinakis

A cascaded finite-state parser for syntactic analysis of Swedish

This report describes the development of a parsing system for written Swedish and is focused on a grammar, the main component of the system, semiautomatically extracted from corpora. A cascaded, finite-state algorithm is applied to the grammar in which the input contains coarse-grained semantic class information, and the output produced reflects not only the syntactic structure of the input, but grammatical functions as well. The grammar has been tested on a variety of random samples of different text genres, achieving precision and recall of 94.62% and 91.92% respectively, and average crossing rate of 0.04, when evaluated against manually disambiguated, annotated texts.

Characteristics of Finnish and Swedish intensive care nursing narratives: a comparative analysis to support the development of clinical language technologies

BackgroundFree text is helpful for entering information into electronic health records, but reusing it is a challenge. The need for language technology for processing Finnish and Swedish healthcare text is therefore evident; however, Finnish and Swedish are linguistically very dissimilar. In this paper we present a comparison of characteristics in Finnish and Swedish free-text nursing narratives from intensive care. This creates a framework for characterising and comparing clinical text and lays the groundwork for developing clinical language technologies.MethodsOur material included daily nursing narratives from one intensive care unit in Finland and one in Sweden. Inclusion criteria for patients were an inpatient period of least five days and an age of at least 16 years. We performed a comparative analysis as part of a collaborative effort between Finnish- and Swedish-speaking healthcare and language technology professionals that included both qualitative and quantitative aspects. The qualitative analysis addressed the content and structure of three average-sized health records from each country. In the quantitative analysis 514 Finnish and 379 Swedish health records were studied using various language technology tools.ResultsAlthough the two languages are not closely related, nursing narratives in Finland and Sweden had many properties in common. Both made use of specialised jargon and their content was very similar. However, many of these characteristics were challenging regarding development of language technology to support producing and using clinical documentation.ConclusionsThe way Finnish and Swedish intensive care nursing was documented, was not country or language dependent, but shared a common context, principles and structural features and even similar vocabulary elements. Technology solutions are therefore likely to be applicable to a wider range of natural languages, but they need linguistic tailoring.AvailabilityThe Finnish and Swedish data can be found at: http://www.dsv.su.se/hexanord/data/.

Fingerprint matrices: uncovering the dynamics of social networks in prose literature

In prose literature often complex dynamics of interpersonal relationships can be observed between the different characters. Traditionally, node‐link diagrams are used to depict the social network of a novel. However, static graphs can only visualize the overall social network structure but not the development of the networks over the course of the story, while dynamic graphs have the serious problem that there are many sudden changes between different portions of the overall social network. In this paper we explore means to show the relationships between the characters of a plot and at the same time their development over the course of a novel. Based on a careful exploration of the design space, we suggest a new visualization technique called Fingerprint Matrices. A case study exemplifies the usage of Fingerprint Matrices and shows that they are an effective means to analyze prose literature with respect to the development of relationships between the different characters.

Anonymisation of Swedish Clinical Data

There is a constantly growing demand for exchanging clinical and health-related information electronically. In the era of the Electronic Health Recordthe release of individual data for research, health care statistics, monitoring of new diagnostic tests and tracking disease outbreak alerts are some of the areas in which the protection of (patient) privacy has become an important concern. In this paper we present a system for automatic anonymisation of Swedish clinical free text, in the form of discharge letters, by applying generic named entity recognition technology.

Mining semantics for culturomics: towards a knowledge-based approach

The massive amounts of text data made available through the Google Books digitization project have inspired a new field of big-data textual research. Named culturomics, this field has attracted the attention of a growing number of scholars over recent years. However, initial studies based on these data have been criticized for not referring to relevant work in linguistics and language technology. This paper provides some ideas, thoughts and first steps towards a new culturomics initiative, based this time on Swedish data, which pursues a more knowledge-based approach than previous work in this emerging field. The amount of new Swedish text produced daily and older texts being digitized in cultural heritage projects grows at an accelerating rate. These volumes of text being available in digital form have grown far beyond the capacity of human readers, leaving automated semantic processing of the texts as the only realistic option for accessing and using the information contained in them. The aim of our recently initiated research program is to advance the state of the art in language technology resources and methods for semantic processing of Big Swedish text and focus on the theoretical and methodological advancement of the state of the art in extracting and correlating information from large volumes of Swedish text using a combination of knowledge-based and statistical methods.

PP-Attachment disambiguation for Swedish : Combining unsupervised and supervised training data

Structural ambiguity, particularly attachment of prepositional phrases, is a serious type of global ambiguity in Natural Language. The disambiguation becomes crucial when a syntactic analyzer must make the correct decision among at least two equally grammatical parse-trees for the same sentence. This paper attempts to find answers to the problem of how attachment ambiguity can be resolved by utilizing Machine Learning (ML) techniques. ML is founded on the assumption that the performance in cognitive tasks is based on the similarity of new situations (testing) to stored representations of earlier experiences (training). Therefore, a large amount of training data is an important prerequisite for providing a solution to the problem. A combination of unsupervised and restricted supervised acquisition of such data will be reported. Training is performed both on a subset of the content of the Gothenburg Lexical Database (GLDB), and on instances of large corpora annotated with coarse-grained semantic information. Testing is performed on corpora instances using a range of different algorithms and metrics. The application language is written Swedish.

Literacy Demands and Information to Cancer Patients

This study examines language complexity of written health information materials for patients undergoing colorectal cancer surgery. Written and printed patient information from 28 Swedish clinics are automatically analyzed by means of language technology. The analysis reveals different problematic issues that might have impact on readability. The study is a first step, and part of a larger project about patients’ health information seeking behavior in relation to written information material. Our study aims to provide support for producing more individualized, person centered information materials according to preferences for complex and detailed or legible texts and thus enhance a movement from receiving information and instructions to participating in knowing. In the near future the study will continue by integrating focus groups with patients that may provide valuable feedback and enhance our knowledge about patients’ use and preferences of different information material.

Shallow Features for Differentiating Disease-Treatment Relations Using Supervised Learning A Pilot Study

Clinical narratives provide an information rich, nearly unexplored corpus of evidential knowledge that is considered as a challenge for practitioners in the language technology field, particularly because of the nature of the texts (excessive use of terminology, abbreviations, orthographic term variation), the significant opportunities for clinical research that such material can provide and the potentially broad impact that clinical findings may have in every day life. It is therefore recognized that the capability to automatically extract key concepts and their relationships from such data will allow systems to properly understand the content and knowledge embedded in the free text which can be of great value for applications such as information extraction and question & answering. This paper gives a brief presentation of such textual data and its semantic annotation, and discusses the set of semantic relations that can be observed between diseases and treatments in the sample. The problem is then designed as a supervised machine learning task in which the relations are tried to be learned using pre-annotated data. The challenges designing the problem and empirical results are presented.

A Graph-Based Analysis of Medical Queries of a Swedish Health Care Portal

Today web portals play an increasingly important role in health care allowing information seekers to learn about diseases and treatments, and to administrate their care. Therefore, it is important that the portals are able to support this process as well as possible. In this paper, we study the search logs of a public Swedish health portal to address the questions if health information seeking differs from other types of Internet search and if there is a potential for utilizing network analysis methods in combination with semantic annotation to gain insights into search behaviors. Using a semantic-based method and a graph-based analysis of word cooccurrences in queries, we show there is an overlap among the results indicating a potential role of these types of methods to gain insights and facilitate improved information search. In addition we show that samples, windows of a month, of search logs may be sufficient to obtain similar results as using larger windows. We also show that medical queries share the same structural properties found for other types of information searches, thereby indicating an ability to reuse existing analysis methods for this type of search data.

Multilingual word embeddings for the assessment of narrative speech in mild cognitive impairment

Abstract We analyze the information content of narrative speech samples from individuals with mild cognitive impairment (MCI), in both English and Swedish, using a combination of supervised and unsupervised learning techniques. We extract information units using topic models trained on word embeddings in monolingual and multilingual spaces, and find that the multilingual approach leads to significantly better classification accuracies than training on the target language alone. In many cases, we find that augmenting the topic model training corpus with additional clinical data from a different language is more effective than training on additional monolingual data from healthy controls. Ultimately we are able to distinguish MCI speakers from healthy older adults with accuracies of up to 63% (English) and 72% (Swedish) on the basis of information content alone. We also compare our method against previous results measuring information content in Alzheimer’s disease, and report an improvement over other topic-modeling approaches. Furthermore, our results support the hypothesis that subtle differences in language can be detected in narrative speech, even at the very early stages of cognitive decline, when scores on screening tools such as the Mini-Mental State Exam are still in the “normal” range.