Physicians are often required to make critical medical decisions that may be based on previous events in the patient's health history. However, these events may be very difficult to locate in the patient record due to the large volume of unstructured textual data in the patient's chart. Even when the chart is housed in an electronic health record (EHR) system, keyword search within the chart may produce many results that are not relevant or that may overlook related expressions and concepts entirely. In addition, some medical events, such as the occurrence of symptoms, are associated with important attributes such as location or severity, and require other elements such as the type of clinical note and its date and time in order to provide the proper context of the event. This paper describes a prototype system that performs ontology-based semantic search through clinical text to extract pain severity events, and then presents them in a visualization to monitor the progression of pain over time.
In this paper, we describe a distant supervision approach for the task of detecting product aspect mentions in customer reviews (e.g., in hotel reviews, we want to associate the aspect ``sleep quality" to a sentence such as "We both slept like rocks."). Detecting such aspects represents an important subtask of aspect-oriented review mining systems, which aim at automatically generating structured summaries of customer opinions. The main advantage of the proposed method is that it allows for the high accuracy of a supervised approach and at the same time avoids the costs of manually labeling a training set. We show how to exploit the inherent structure of customer reviews to automatically gather large amounts of labeled data. Our experimental results show that the method achieves a performance as good as a traditional, fully supervised approach.
Expression of the gene, OsAPX4, coding for ascorbate peroxidase in leaves and roots of rice were induced by abiotic stresses, such as NaCl, NaHCO3 and Na2CO3, polyethylene glycol (PEG) 6000, H2O2, CuCl2. Yeast (Saccharomyces cerevisiae) over-expressing ascorbate peroxidase exhibited greater tolerance to NaCl and NaHCO3 and transgenic Arabidopsis over-expressing OsAPX4 had a greater salt tolerance than wild-type plants in 1/2 Murashige and Skoog (MS) medium with 150, 200 mM NaCl and 5, 7.5 mM NaHCO3. These results suggest that OsAPX4 plays an important role in multiple environmental stresses.
This paper describes a new approach for estimating term weights in a document, and shows how the new weighting scheme can be used to improve the accuracy of a text classifier. The method uses term co-occurrence as a measure of dependency between word features. A random walk model is applied on a graph encoding words and co-occurrence dependencies, resulting in scores that represent a quantification of how a particular word feature contributes to a given context. Experiments performed on three standard classification datasets show that the new random walk based approach outperforms the traditional term frequency approach of feature weighting.
Mobile agent technology has a number of characteristics that make it well suited for distributed information retrieval and monitoring tasks. We present an agent-based environment that provides facilities and tools for remote diagnosis; supervision and control of process automation systems. The environment allows remote and dynamic configuration of diagnosis and supervision tasks,offers a framework for observing and controlling existing (legacy) software, and is also suited for other application domains than process automation.
This paper describes a real world application of Data Mining methods for deviation detection. The goal is to analyse warranty claims in the automobile sector. Basically we want to support the technical engineers concerned with warranty issues in two ways: First of all we want to guide them during verification of their hypothesis and additionally Ne want to strengthen their creative and inspirational potentials. For this purpose we accessed the Quality Information System (QUIS) of DaimlerChrysler. The whole project was carried through according to the CRISP-DM data mining process. The methods from Data Mining that we applied were: baysian nets, boolean association rules, generalised association rules, quantitative association rules and sequential patterns. We present some of the data mining results exemplarily, discuss the difficulties we encountered and finally give a short conclusion.
This paper presents the design and implementation methodology of the JCQ system, a Java-based Continual Query system for update monitoring over Web information sources. A continual query is a standing query that monitors updates of interest using distributed triggers and notifies users whenever the updates reach specified thresholds. In this paper we focus on the strategies and techniques developed in JCQ for scalable and efficient trigger firing and the execution model for flexible and robust change notification. We evaluate our approach through a performance study of the most recent release of the JCQ system and a comparison with related work.
