Big Data on Healthcare Application

Permanent URI for this collection

Browse

Recent Submissions

Now showing 1 - 5 of 7
  • Item
    High-performance detection of alcoholism by unfolding the amalgamated EEG spectra using the Random Forests method
    ( 2019-01-08) Rieg, Thilo ; Frick, Janek ; Hitzler, Marius ; Buettner, Ricardo
    We show that by unfolding the outdated EEG standard bandwidths in a fine-grade equidistant 99-point spectrum we can precisely detect alcoholism. Using this novel pre-processing step prior to entering a random forests classifier, our method substantially outperforms all previous results with a balanced accuracy of 97.4 percent. Our machine learning work contributes to healthcare and information systems. Due to its drastic and protracted consequences, alcohol consumption is always a critical issue in our society. Consequences of alcoholism in the brain can be recorded using electroencephalography (EEG). Our work can be used to automatically detect alcoholism in EEG mass data within milliseconds. In addition, our results challenge the medically outdated EEG standard bandwidths.
  • Item
    Acute Respiratory and Cardiovascular Outcomes Associated with Low Levels of Ambient Fine Particulate Matter (PM2.5) on the Island of Oahu
    ( 2019-01-08) Mnatzaganian, Christina ; Lozano, Alicia ; Pellegrin, Karen ; Miyamura, Jill ; Knox, Mark ; Hanlon, Alexandra
    Scant literature exists regarding health effects of fine particulate matter (PM2.5) pollution at or below national standards. This study examined the relationship between PM2.5 and acute care use and costs in Honolulu where PM2.5 is low. Single and distributed lag over-dispersed Poisson models were used to examine hospitalizations/emergency department (ED) visits associated with cumulative PM2.5 exposure over the current day and seven previous days (lags 0-7) in 2011. A 10-µg/m3 increase in cumulative PM2.5 concentration was associated with a 32% increase in respiratory admissions (RR=1.32, p=0.001) costing $486,908 and a 24% decrease in respiratory admissions in the comparison group (RR=0.76, p<0.001). ED visits increased by 12% at lag day 0 for respiratory outcomes (RR=1.12, p=0.03) and cumulatively with increased respiratory visits by 49% (RR=1.49) and increased combined respiratory and cardiovascular issues by 20% (RR=1.20; p<0.01 for both) costing $117,856. Additional research is needed on health effects within pollution lower levels.
  • Item
    Mining and Representing Unstructured Nicotine Use Data in a Structured Format for Secondary Use
    ( 2019-01-08) Ngwenya, Mandlenkosi ; Bankole, Felix
    The objective of this study was to use rules, NLP and machine learning for addressing the problem of clinical data interoperability across healthcare providers. Addressing this problem has the potential to make clinical data comparable, retrievable and exchangeable between healthcare providers. Our focus was in giving structure to unstructured patient smoking information. We collected our data from the MIMIC-III database. We wrote rules for annotating the data, then trained a CRF sequence classifier. We obtained an f-measure of 86%, 72%, 69%, 80%, and 12% for substance smoked, frequency, amount, temporal, and duration respectively. Amount smoked yielded a small value due to scarcity of related data. Then for smoking status we obtained an f-measure of 94.8% for non-smoker class, 83.0% for current-smoker, and 65.7% for past-smoker. We created a FHIR profile for mapping the extracted data based on openEHR reference models, however in future we will explore mapping to CIMI models.
  • Item
    A Novel Model for Classification of Parkinson’s Disease: Accurately Identifying Patients for Surgical Therapy
    ( 2019-01-08) Mohammed, Farhan ; He, Xiangjian ; Lin, Yiguang ; Chen, Jinjun
    Parkinson’s disease (PD) is a neurodegenerative disorder and a global health problem that has no curative therapies. Surgery is a well-established therapy for controlling symptoms of advanced PD patients. This paper proposes a streamlined model to classify PD and to identify appropriate patients for surgical therapy. The data was gathered from the Parkinson's Progressive Markers Initiative consisting of 1080 subjects. Multilayer Perceptron (MLP), Decision trees, Support Vector Machine and Naïve Bayes are used as classifiers. MLP achieves the highest accuracy as compared to other three classifiers. The dataset used in our experiments is from the Parkinson Progressive Markers Initiative. With feature selection, it is observed that the same classification accuracy is achieved with 60% of the attributes as that using all attributes. It is demonstrated that our classification model for PD patients produces the most accurate results and achieves the highest accuracy of 98.13%.
  • Item