De-Identification of health records using Anonym: Effectiveness and robustness across datasets

Objective Evaluate the effectiveness and robustness of Anonym, a tool for de-identifying free-text health records based on conditional random fields classifiers informed by linguistic and lexical features, as well as features extracted by pattern matching techniques. De-identification of personal health information in electronic health records is essential for the sharing and secondary usage of clinical data. De-identification tools that adapt to different sources of clinical data are attractive as they would require minimal intervention to guarantee high effectiveness. Methods and Materials The effectiveness and robustness of Anonym are evaluated across multiple datasets, including the widely adopted Integrating Biology and the Bedside (i2b2) dataset, used for evaluation in a de-identification challenge. The datasets used here vary in type of health records, source of data, and their quality, with one of the datasets containing optical character recognition errors. Results Anonym identifies and removes up to 96.6% of personal health identifiers (recall) with a precision of up to 98.2% on the i2b2 dataset, outperforming the best system proposed in the i2b2 challenge. The effectiveness of Anonym across datasets is found to depend on the amount of information available for training. Conclusion Findings show that Anonym compares to the best approach from the 2006 i2b2 shared task. It is easy to retrain Anonym with new datasets; if retrained, the system is robust to variations of training size, data type and quality in presence of sufficient training data.

Crowdsourcing interactions : using crowdsourcing for evaluating interactive information retrieval systems

In the field of information retrieval (IR), researchers and practitioners are often faced with a demand for valid approaches to evaluate the performance of retrieval systems. The Cranfield experiment paradigm has been dominant for the in-vitro evaluation of IR systems. Alternative to this paradigm, laboratory-based user studies have been widely used to evaluate interactive information retrieval (IIR) systems, and at the same time investigate users’ information searching behaviours. Major drawbacks of laboratory-based user studies for evaluating IIR systems include the high monetary and temporal costs involved in setting up and running those experiments, the lack of heterogeneity amongst the user population and the limited scale of the experiments, which usually involve a relatively restricted set of users. In this paper, we propose an alternative experimental methodology to laboratory-based user studies. Our novel experimental methodology uses a crowdsourcing platform as a means of engaging study participants. Through crowdsourcing, our experimental methodology can capture user interactions and searching behaviours at a lower cost, with more data, and within a shorter period than traditional laboratory-based user studies, and therefore can be used to assess the performances of IIR systems. In this article, we show the characteristic differences of our approach with respect to traditional IIR experimental and evaluation procedures. We also perform a use case study comparing crowdsourcing-based evaluation with laboratory-based evaluation of IIR systems, which can serve as a tutorial for setting up crowdsourcing-based IIR evaluations.

Automated classification of limb fractures from free-text radiology reports using a clinician-informed gazetteer methodology

Background Timely diagnosis and reporting of patient symptoms in hospital emergency departments (ED) is a critical component of health services delivery. However, due to dispersed information resources and a vast amount of manual processing of unstructured information, accurate point-of-care diagnosis is often difficult. Aims The aim of this research is to report initial experimental evaluation of a clinician-informed automated method for the issue of initial misdiagnoses associated with delayed receipt of unstructured radiology reports. Method A method was developed that resembles clinical reasoning for identifying limb abnormalities. The method consists of a gazetteer of keywords related to radiological findings; the method classifies an X-ray report as abnormal if it contains evidence contained in the gazetteer. A set of 99 narrative reports of radiological findings was sourced from a tertiary hospital. Reports were manually assessed by two clinicians and discrepancies were validated by a third expert ED clinician; the final manual classification generated by the expert ED clinician was used as ground truth to empirically evaluate the approach. Results The automated method that attempts to individuate limb abnormalities by searching for keywords expressed by clinicians achieved an F-measure of 0.80 and an accuracy of 0.80. Conclusion While the automated clinician-driven method achieved promising performances, a number of avenues for improvement were identified using advanced natural language processing (NLP) and machine learning techniques.

Classification of cancer-related death certificates using machine learning

Background Cancer monitoring and prevention relies on the critical aspect of timely notification of cancer cases. However, the abstraction and classification of cancer from the free-text of pathology reports and other relevant documents, such as death certificates, exist as complex and time-consuming activities. Aims In this paper, approaches for the automatic detection of notifiable cancer cases as the cause of death from free-text death certificates supplied to Cancer Registries are investigated. Method A number of machine learning classifiers were studied. Features were extracted using natural language techniques and the Medtex toolkit. The numerous features encompassed stemmed words, bi-grams, and concepts from the SNOMED CT medical terminology. The baseline consisted of a keyword spotter using keywords extracted from the long description of ICD-10 cancer related codes. Results Death certificates with notifiable cancer listed as the cause of death can be effectively identified with the methods studied in this paper. A Support Vector Machine (SVM) classifier achieved best performance with an overall F-measure of 0.9866 when evaluated on a set of 5,000 free-text death certificates using the token stem feature set. The SNOMED CT concept plus token stem feature set reached the lowest variance (0.0032) and false negative rate (0.0297) while achieving an F-measure of 0.9864. The SVM classifier accounts for the first 18 of the top 40 evaluated runs, and entails the most robust classifier with a variance of 0.001141, half the variance of the other classifiers. Conclusion The selection of features significantly produced the most influences on the performance of the classifiers, although the type of classifier employed also affects performance. In contrast, the feature weighting schema created a negligible effect on performance. Specifically, it is found that stemmed tokens with or without SNOMED CT concepts create the most effective feature when combined with an SVM classifier.

Understanding negation and family history to improve clinical information retrieval

We present a study to understand the effect that negated terms (e.g., "no fever") and family history (e.g., "family history of diabetes") have on searching clinical records. Our analysis is aimed at devising the most effective means of handling negation and family history. In doing so, we explicitly represent a clinical record according to its different content types: negated, family history and normal content; the retrieval model weights each of these separately. Empirical evaluation shows that overall the presence of negation harms retrieval effectiveness while family history has little effect. We show negation is best handled by weighting negated content (rather than the common practise of removing or replacing it). However, we also show that many queries benefit from the inclusion of negated content and that negation is optimally handled on a per-query basis. Additional evaluation shows that adaptive handing of negated and family history content can have significant benefits.

Relevation! : an open source system for information retrieval relevance assessment

Relevation! is a system for performing relevance judgements for information retrieval evaluation. Relevation! is web-based, fully configurable and expandable; it allows researchers to effectively collect assessments and additional qualitative data. The system is easily deployed allowing assessors to smoothly perform their relevance judging tasks, even remotely. Relevation! is available as an open source project at: http://ielab.github.io/relevation.

Top-k retrieval using facility location analysis

The top-k retrieval problem aims to find the optimal set of k documents from a number of relevant documents given the user’s query. The key issue is to balance the relevance and diversity of the top-k search results. In this paper, we address this problem using Facility Location Analysis taken from Operations Research, where the locations of facilities are optimally chosen according to some criteria. We show how this analysis technique is a generalization of state-of-the-art retrieval models for diversification (such as the Modern Portfolio Theory for Information Retrieval), which treat the top-k search results like “obnoxious facilities” that should be dispersed as far as possible from each other. However, Facility Location Analysis suggests that the top-k search results could be treated like “desirable facilities” to be placed as close as possible to their customers. This leads to a new top-k retrieval model where the best representatives of the relevant documents are selected. In a series of experiments conducted on two TREC diversity collections, we show that significant improvements can be made over the current state-of-the-art through this alternative treatment of the top-k retrieval problem.

A comprehensive analysis of parameter settings for novelty-biased cumulative gain

In the TREC Web Diversity track, novelty-biased cumulative gain (α-NDCG) is one of the official measures to assess retrieval performance of IR systems. The measure is characterised by a parameter, α, the effect of which has not been thoroughly investigated. We find that common settings of α, i.e. α=0.5, may prevent the measure from behaving as desired when evaluating result diversification. This is because it excessively penalises systems that cover many intents while it rewards those that redundantly cover only few intents. This issue is crucial since it highly influences systems at top ranks. We revisit our previously proposed threshold, suggesting α be set on a query-basis. The intuitiveness of the measure is then studied by examining actual rankings from TREC 09-10 Web track submissions. By varying α according to our query-based threshold, the discriminative power of α-NDCG is not harmed and in fact, our approach improves α-NDCG's robustness. Experimental results show that the threshold for α can turn the measure to be more intuitive than using its common settings.

A query-basis approach to parametrizing novelty-biased cumulative gain

Novelty-biased cumulative gain (α-NDCG) has become the de facto measure within the information retrieval (IR) community for evaluating retrieval systems in the context of sub-topic retrieval. Setting the incorrect value of parameter α in α-NDCG prevents the measure from behaving as desired in particular circumstances. In fact, when α is set according to common practice (i.e. α = 0.5), the measure favours systems that promote redundant relevant sub-topics rather than provide novel relevant ones. Recognising this characteristic of the measure is important because it affects the comparison and the ranking of retrieval systems. We propose an approach to overcome this problem by defining a safe threshold for the value of α on a query basis. Moreover, we study its impact on system rankings through a comprehensive simulation.

Investigating subspace distances in semantic spaces

Semantic space models of word meaning derived from co-occurrence statistics within a corpus of documents, such as the Hyperspace Analogous to Language (HAL) model, have been proposed in the past. While word similarity can be computed using these models, it is not clear how semantic spaces derived from different sets of documents can be compared. In this paper, we focus on this problem, and we revisit the proposal of using semantic subspace distance measurements [1]. In particular, we outline the research questions that still need to be addressed to investigate and validate these distance measures. Then, we describe our plans for future research.

An analysis of ranking principles and retrieval strategies

The assumptions underlying the Probability Ranking Principle (PRP) have led to a number of alternative approaches that cater or compensate for the PRP’s limitations. All alternatives deviate from the PRP by incorporating dependencies. This results in a re-ranking that promotes or demotes documents depending upon their relationship with the documents that have been already ranked. In this paper, we compare and contrast the behaviour of state-of-the-art ranking strategies and principles. To do so, we tease out analytical relationships between the ranking approaches and we investigate the document kinematics to visualise the effects of the different approaches on document ranking.

When two is better than one : a study of ranking paradigms and their integrations for subtopic retrieval

In this paper, we consider the problem of document ranking in a non-traditional retrieval task, called subtopic retrieval. This task involves promoting relevant documents that cover many subtopics of a query at early ranks, providing thus diversity within the ranking. In the past years, several approaches have been proposed to diversify retrieval results. These approaches can be classified into two main paradigms, depending upon how the ranks of documents are revised for promoting diversity. In the first approach subtopic diversification is achieved implicitly, by choosing documents that are different from each other, while in the second approach this is done explicitly, by estimating the subtopics covered by documents. Within this context, we compare methods belonging to the two paradigms. Furthermore, we investigate possible strategies for integrating the two paradigms with the aim of formulating a new ranking method for subtopic retrieval. We conduct a number of experiments to empirically validate and contrast the state-of-the-art approaches as well as instantiations of our integration approach. The results show that the integration approach outperforms state-of-the-art strategies with respect to a number of measures.

University of Glasgow at ImageCLEFPhoto 2009 : optimising similarity and diversity in image retrieval

In this paper we describe the approaches adopted to generate the runs submitted to ImageCLEFPhoto 2009 with an aim to promote document diversity in the rankings. Four of our runs are text based approaches that employ textual statistics extracted from the captions of images, i.e. MMR [1] as a state of the art method for result diversification, two approaches that combine relevance information and clustering techniques, and an instantiation of Quantum Probability Ranking Principle. The fifth run exploits visual features of the provided images to re-rank the initial results by means of Factor Analysis. The results reveal that our methods based on only text captions consistently improve the performance of the respective baselines, while the approach that combines visual features with textual statistics shows lower levels of improvements.

The Quantum Probability Ranking Principle for Information Retrieval

While the Probability Ranking Principle for Information Retrieval provides the basis for formal models, it makes a very strong assumption regarding the dependence between documents. However, it has been observed that in real situations this assumption does not always hold. In this paper we propose a reformulation of the Probability Ranking Principle based on quantum theory. Quantum probability theory naturally includes interference effects between events. We posit that this interference captures the dependency between the judgement of document relevance. The outcome is a more sophisticated principle, the Quantum Probability Ranking Principle, that provides a more sensitive ranking which caters for interference/dependence between documents’ relevance.

Semantic spaces : measuring the distance between different subspaces

Semantic Space models, which provide a numerical representation of words’ meaning extracted from corpus of documents, have been formalized in terms of Hermitian operators over real valued Hilbert spaces by Bruza et al. [1]. The collapse of a word into a particular meaning has been investigated applying the notion of quantum collapse of superpositional states [2]. While the semantic association between words in a Semantic Space can be computed by means of the Minkowski distance [3] or the cosine of the angle between the vector representation of each pair of words, a new procedure is needed in order to establish relations between two or more Semantic Spaces. We address the question: how can the distance between different Semantic Spaces be computed? By representing each Semantic Space as a subspace of a more general Hilbert space, the relationship between Semantic Spaces can be computed by means of the subspace distance. Such distance needs to take into account the difference in the dimensions between subspaces. The availability of a distance for comparing different Semantic Subspaces would enable to achieve a deeper understanding about the geometry of Semantic Spaces which would possibly translate into better effectiveness in Information Retrieval tasks.