Life expectancy is one of the most important factors in end-of-life decision making. Good prognostication for example helps to determine the course of treatment and helps to anticipate the procurement of health care services and facilities, or more broadly: facilitates Advance Care Planning. Advance Care Planning improves the quality of the final phase of life by stimulating doctors to explore the preferences for end-of-life care with their patients, and people close to the patients. Physicians, however, tend to overestimate life expectancy, and miss the window of opportunity to initiate Advance Care Planning. This research tests the potential of using machine learning and natural language processing techniques for predicting life expectancy from electronic medical records. We approached the task of predicting life expectancy as a supervised machine learning task. We trained and tested a long short-term memory recurrent neural network on the medical records of deceased patients. We developed the model with a ten-fold cross-validation procedure, and evaluated its performance on a held-out set of test data. We compared the performance of a model which does not use text features (baseline model) to the performance of a model which uses features extracted from the free texts of the medical records (keyword model), and to doctors’ performance on a similar task as described in scientific literature. Both doctors and the baseline model were correct in 20% of the cases, taking a margin of 33% around the actual life expectancy as the target. The keyword model, in comparison, attained an accuracy of 29% with its prognoses. While doctors overestimated life expectancy in 63% of the incorrect prognoses, which harms anticipation to appropriate end-of-life care, the keyword model overestimated life expectancy in only 31% of the incorrect prognoses. Prognostication of life expectancy is difficult for humans. Our research shows that machine learning and natural language processing techniques offer a feasible and promising approach to predicting life expectancy. The research has potential for real-life applications, such as supporting timely recognition of the right moment to start Advance Care Planning.

Talking about odors and flavors is difficult for most people, yet experts appear to be able to convey critical information about wines in their reviews. This seems to be a contradiction, and wine expert descriptions are frequently received with criticism. Here, we propose a method for probing the language of wine reviews, and thus offer a means to enhance current vocabularies, and as a by-product question the general assumption that wine reviews are gibberish. By means of two different quantitative analyses—support vector machines for classification and Termhood analysis—on a corpus of online wine reviews, we tested whether wine reviews are written in a consistent manner, and thus may be considered informative; and whether reviews feature domain-specific language. First, a classification paradigm was trained on wine reviews from one set of authors for which the color, grape variety, and origin of a wine were known, and subsequently tested on data from a new author. This analysis revealed that, regardless of individual differences in vocabulary preferences, color and grape variety were predicted with high accuracy. Second, using Termhood as a measure of how words are used in wine reviews in a domain-specific manner compared to other genres in English, a list of 146 wine-specific terms was uncovered. These words were compared to existing lists of wine vocabulary that are currently used to train experts. Some overlap was observed, but there were also gaps revealed in the extant lists, suggesting these lists could be improved by our automatic analysis.

Various Machine Learning (ML) approaches have been demonstrated to produce relatively successful Word Sense Disambiguation (WSD) systems. There are still unexplained differences among the performance measurements of different algorithms, hence it is warranted to deepen the investigation into which algorithm has the right ‘bias’ for this task. In this paper, we show that this is not easy to accomplish, due to intricate interactions between information sources, parameter settings, and properties of the training data. We investigate the impact of parameter optimization on generalization accuracy in a memory-based learning approach to English and Dutch WSD. A ‘word-expert’ architecture was adopted, yielding a set of classifiers, each specialized in one single wordform. The experts consist of multiple memory-based learning classifiers, each taking different information sources as input, combined in a voting scheme. We optimized the architectural and parametric settings for each individual word-expert by performing cross-validation experiments on the learning material. The results of these experiments show that the variation of both the algorithmic parameters and the information sources available to the classifiers leads to large fluctuations in accuracy. We demonstrate that optimization per word-expert leads to an overall significant improvement in the generalization accuracies of the produced WSD systems.

Background: Regulatory bodies such as health care inspectorates can identify potential patient safety problems in health care providers by analyzing patient complaints. However, it is challenging to analyze the large number of complaints. Text mining techniques may help identify signals of problems with patient safety at health care providers. Objective: The aim of this study was to explore whether employing text mining techniques on patient complaint databases can help identify potential problems with patient safety at health care providers and automatically predict the severity of patient complaints. Methods: We performed an exploratory study on the complaints database of the Dutch Health and Youth Care Inspectorate with more than 22,000 written complaints. Severe complaints are defined as those cases where the inspectorate contact point experts deemed it worthy of a triage by the inspectorate, or complaints that led to direct action by the inspectorate. We investigated a range of supervised machine learning techniques to assign a severity label to complaints that can be used to prioritize which incoming complaints need the most attention. We studied several features based on the complaints’ written content, including sentiment analysis, to decide which were helpful for severity prediction. Finally, we showcased how we could combine these severity predictions and automatic keyword analysis on the complaints database and listed health care providers and their organization-specific complaints to determine the average severity of complaints per organization. Results: A straightforward text classification approach using a bag-of-words feature representation worked best for the severity prediction of complaints. We obtained an accuracy of 87%-93% (2658-2990 of 3319 complaints) on the held-out test set and an F1 score of 45%-51% on the severe complaints. The skewed class distribution led to only reasonable recall (47%-54%) and precision (44%-49%) scores. The use of sentiment analysis for severity prediction was not helpful. By combining the predicted severity outcomes with an automatic keyword analysis, we identified several health care providers that could have patient safety problems. Conclusions: Text mining techniques for analyzing complaints by civilians can support inspectorates. They can automatically predict the severity of the complaints, or they can be used for keyword analysis. This can help the inspectorate detect potential patient safety problems, or support prioritizing follow-up supervision activities by sorting complaints based on the severity per organization or per sector.