MACHINE LEARNING AND WEARABLE DEVICES FOR PHONOCARDIOGRAM-BASED DIAGNOSIS
Shaima Abdelmageed and Mohammed Elmusrati, University of Vasa, Finland
The heart sound signal, Phonocardiogram (PCG) is difficult to interpret even for experienced cardiologists. Interpretation are very subjective depending on the hearing ability of the physician. mHealth has been the adopted approach towards simplifying that and getting quick diagnosis using mobile devices. However, it has been challenging due to the required high quality of data, high computation load, and high-power consumption. The aim of this paper is to diagnose the heart condition based on Phonocardiogram analysis using Machine Learning techniques assuming limited processing power to be encapsulated later in a wearable device. The cardiovascular system is modelled in a transfer function to provide PCG signal recording as it would be recorded at the wrist. The signal is, then, decomposed using filter bank and the analysed using discriminant function. The results showed that PCG with a 19 dB Signal-to-Noise-Ratio can lead to 97.33% successful diagnosis. The same decomposed signal is then analysed using pattern recognition neural network, and the classification was 100% successful with 83.3% trust level.
Analysis, Classification, data quality, diagnosis, filter banks, mHealth, PCG, SNR, transfer function, Wavelet Transform, wearable
CONSTRUCTING A SEMANTIC GRAPH WITH DEPRESSION SYMPTOMS EXTRACTED FROM TWITTER
Long Ma, Troy University, USA
Depression diagnosis is a critical challenge in mental precision medicine since there is currently no a gold standard using depression symptoms. Usually, a doctor makes depression diagnosis based on patients’ answers to interview questions. The depression diagnosis depends on a person’s behavior symptoms. Due to privacy of clinical data in a hospital, it is very hard to get patients’ medical data. Thus, we directly use public social media data containing much information from patients, doctors and other people on Twitter. The research goal is to extract depression symptoms from the massive social data from Twitter via text mining and then make a semantic graph to representing relations among the depression symptoms. Different from commonly used statistical methods, we propose a hybrid method that integrates the statistical analysis and natural language processing techniques to make the semantic graph with the discovered depression symptoms from tweets. In the future, the depression symptom semantic graph will be used to build an intelligent depression diagnosis software system for medical doctors and a convenient depression self-screening software system for ordinary people.
Depression, Social Media, Text Mining, Twitter, Social Networks, Natural Language Processing, Word2Vec