Predicting treatment outcome and drug adverse effects by temporal data mining and interactive visualization | |
| Author | Wipada Chanthaweethip |
| Call Number | AIT Thesis no.IM-12-06 |
| Subject(s) | Data mining--Thailand Visualization HIV infections--Thailand |
| Note | A thesis submitted in partial fulfillment of the requirements for the degree of Master of Science in Information Management, School of Engineering and Technology |
| Publisher | Asian Institute of Technology |
| Series Statement | Thesis ; no. IM-12-06 |
| Abstract | Number of HIV patients who need treatment is increasing continuously, 36% of them are treated with antiretroviral (ARV) treatment. In resource limited setting, treating patients with ARV is to maintain them in the first line regimen as long as possible. Therefore, the events which could interrupt the treatment should be avoided. Early detecting those events would be a solution for maintaining patients in their first regimen. This study proposes treatment outcome and drug adverse effects prediction by using temporal data mining and integrated with interactive visualization. The purpose is to assist non-specialized physician and reduce work load of monitoring with large number of patients. Temporal abstraction is used for classifying time series data into discrete categories, each represented typically with a symbol. Artificial Neural Network is employed while problem of imbalance dataset is occurred in learning part and it reduces prediction performance. The proposed solution is to resampling data by under-sampling them. Two resampling and original dataset were used in model training and testing where the accuracy from those three dataset are compared. Visualization was used for representing prediction result and provide user interaction. The result shows that the prediction model can produce more than 85% of accuracy with resampling 2 dataset. The variables which have significant weight in predicting both tasks are associated with real world diagnosis. VL is primary feature for predict treatment outcome while Cholesterol, Triglyceride, SGPT, SGOT, and eGFR are primary features for predict adverse effects. |
| Year | 2012 |
| Corresponding Series Added Entry | Asian Institute of Technology. Thesis ; no. IM-12-06 |
| Type | Thesis |
| School | School of Engineering and Technology |
| Department | Department of Information and Communications Technologies (DICT) |
| Academic Program/FoS | Information Management (IM) |
| Chairperson(s) | Guha, Sumanta ; |
| Examination Committee(s) | Vatcharaporn Esichaikul;Duboz, Raphael ; |
| Scholarship Donor(s) | Royal Thai Government Fellowship; |
| Degree | Thesis (M.Sc.) - Asian Institute of Technology, 2012 |