Developing the Predictive Model for Antiretroviral Therapy Adherence Status of Human Immunodeficiency Virus Patients Care Using Deep Neural Network

Mitiku, Melkamu

dc.contributor.author	Mitiku, Melkamu
dc.date.accessioned	2020-10-07T11:29:05Z
dc.date.available	2020-10-07T11:29:05Z
dc.date.issued	2020
dc.identifier.uri	http://hdl.handle.net/123456789/11288
dc.description.abstract	Now a day’s health care providing facilities, organizations, and institutes generate vast amounts of data on their business transaction. Models or systems should be designed and developed by using these vast amounts of data in to help and make lighten their health care service provision system. AI is a field of computer science by which intelligent system is designed which mimic human brain. ML is subfield of AI which is pattern recognition process using dataset as input and algorithms. A DNN is subfield of ML, which are used for designing a predictive model using supervised machine learning techniques. In health care organization, great problem is faced health care providers to know the ART adherence status of HIV/AIDS patients. In this research, ART adherence status predictive model using DNN have been developed to let service providers know the ART adherence status of HIV/AIDS patients using some features of the patient’s treatments; to make easy their work in HIV/AIDS patients care and treatment process. ART adherence predictive model is designed in the process of dataset collection, dataset preprocessing like missing data, outlier data and incomplete data treatment, feature extraction and feature transformation and normalization the dataset. The dataset has been split into training, validation and testing dataset i.e. 70% for training and validating the algorithm and 30% for testing the model. Model designing, model training, evaluation and finally testing was performed using anaconda/miniconda, keras tensorflow and python programing language. The model developed performed at training accuracy level of 97.17% and a mean square error of 0.03. For future recommendation any interested researcher can improve the quality of this model using additional key feature like OI, drug toxicity, co-infection with related to ARV, life style and other feature of the HIV/AIDS patients.	en_US
dc.language.iso	en	en_US
dc.subject	Information Technology	en_US
dc.title	Developing the Predictive Model for Antiretroviral Therapy Adherence Status of Human Immunodeficiency Virus Patients Care Using Deep Neural Network	en_US
dc.type	Thesis	en_US