عنوان
پدید آورنده
موضوع
رده
کتابخانه
محل استقرار
TLpq2288064641
انگلیسی
Social Media Mining Algorithms Big Data Analysts Need to Improve Infectious Disease Predictive Model for Malaria in Ethiopia
[Thesis]
Tsehay, Yoseph
Schmitt, Alexa
Colorado Technical University
2019
155
D.C.S.
Colorado Technical University
2019
Understanding big data and data mining algorithms is vital in the study of predictive analytics and the healthcare industry in improving disease predictive models. Disease predictive models can predict the probabilities of infectious disease outbreaks in advance by determining the dependency of a disease outbreak on the risk factors of the disease. However, the performance of the predictive model is dependent on the availability of risk factor data. In this research, the poor performance of malaria predictive algorithms due to lack of disease risk factor data was addressed. A hybrid of social media mining algorithms was used to generate additional data from social media to improve a malaria predictive model constructed for Ethiopia. A design science research method was used to identify the problem, define solution objectives, design and develop, and demonstrate and evaluate the proposed solution. The demonstration of a combined dataset extracted from Facebook, malaria incidence reports, and climate data showed that a Bayesian Belief Network predictive model is 8% more accurate than traditional predictive models and can provide uncertainty estimates when some data are unavailable. Subject matter experts reviewed the Bayesian Network and their review was incorporated for further improvement of the predictive model. Additionally, the experts were in complete agreement on the substantial role of social media mining algorithms in improving infectious disease predictive models. This study adds to the body of knowledge in areas of social media mining, predictive analytics, and infectious disease predictive modeling.
Computer science
Schmitt, Alexa
Tsehay, Yoseph
مطالعه متن کتاب p
[Thesis]
276903
a
Y
