Big Data Analytics Approach for E-health: The Studies of Memory, Attention Capacities and Sleep Disorder
| Autor: | Dinh-Van Phan, 潘庭問 |
|---|---|
| Rok vydání: | 2018 |
| Druh dokumentu: | 學位論文 ; thesis |
| Popis: | 107 Big data analysis has been creating opportunities and challenges for many fields in our lives such as science, business, medical, healthcare, etc. Thus, if we know big data analysis techniques, we may create advantages in business, add more values in research, etc. This dissertation focused on the analysis of the big data in health care and disease diagnosis. It surveys the effect of daily physical activities as well as daily sleep on memory and attention capacities. This dissertation also studied asthma patient cohort from 2001-2010 in Taiwan base on NHIRD. Eventually, it applied traditional machine learning algorithms and deep learning to sleep disorder prediction in new asthma cohort from 2002-2010. This dissertation indicated associations between daily physical activities, daily sleep and memory, attention capacity. The results show that there are significant negative correlations between memory capacity, time spent on the attention test (TSAT) and very active time duration (VATD), calories burnt on the day before the test date (r = 0 289, r = 0 254, r = −0 272, r = −0 176, respectively) and during the week before the test date (r = 0 268, r = 0 241, r = −0 364, r = −0 395, respectively). In addition, the thresholds to best discriminate between normal-to-good and low attention capacity of the VATD and the calories burnt per day were ≥2283 calories, ≥20 minutes on the day before testing, or ≥13,640 calories per week, ≥76 minutes per week during the week before the test date. The findings indicate the short-term effects that VATD and calories burnt on the day before or during the week before the test date significantly and negatively associated with memory and attention capacities of college students. This study also indicates a significant negative correlation between memory capacity and awake count on the test date and during the week before the test date (r = −0.153, r = −0.391, respectively). However, the minutes asleep on the test date and during the week before the test date significantly positively associates with memory capacity (r = 0.127, r = 0.370, respectively). Furthermore, spending > 6 hours and 42 minutes asleep on the test date or > 6 hours and 37 minutes asleep per day on average during the week before the test date result in better memory capacity. Average annual percentage changes (AAPCs) in asthma cohort has shown that the number of asthma patients visiting emergency (ED) and non-emergency (non-ED) clinics significantly increased (2.3% and 4.6%, respectively) from 2001-2010. The asthma visits classified by hospital level showed that the local hospitals and others exhibited a significant increasing trend (AAPC=15.3%). The annual prevalence of children group was the highest significantly increased (AAPC=3.9%). Three widely-used machine learning algorithms K-Nearest Neighbors, Support Vector Machine, Random forest, and deep learning models were applied to the new asthma cohort from 2002-2010 for predicting sleep disorder. The results show that CNN models achieve the highest accuracy (>92.3%). |
| Databáze: | Networked Digital Library of Theses & Dissertations |
| Externí odkaz: |
|