Risk analysis of prostate cancer occurrence based on PLCO data
Principal Investigator
Name
Lukasz Markiewicz
Degrees
M.Sc.
Institution
Independent
Position Title
Postgraduate student
Email
About this CDAS Project
Study
PLCO
(Learn more about this study)
Project ID
PLCO-332
Initial CDAS Request Approval
Jan 17, 2018
Title
Risk analysis of prostate cancer occurrence based on PLCO data
Summary
The main goal of my project is to perform statistical analysis of prostate cancer datasets and build a model which would help to assess the risk of cancer occurrence.
Additionally since prostate cancer seems to be less frequent in South-East Asia (China especially) than in Western countries and the main obvious difference between both areas are dietary habits, I would like to test if I can identify any significant correlation in dietary and diet history data sets compared to the main prostate cancer data set.
Also using ancillary data I would like to verify existence of any potential trends that might occur since patients had started their participation in the study. As PSA level measurement was also collected during the study and it is a commonly used indicator in prostate cancer diagnostics, I would like to include this variable as precisely as I can in a potential model that will arise in my project. Controversy that is associated with PSA level related diagnostic seems to be even more interesting when comparing to any other correlation that can be identified based on the prostate cancer dataset.
Additionally since prostate cancer seems to be less frequent in South-East Asia (China especially) than in Western countries and the main obvious difference between both areas are dietary habits, I would like to test if I can identify any significant correlation in dietary and diet history data sets compared to the main prostate cancer data set.
Also using ancillary data I would like to verify existence of any potential trends that might occur since patients had started their participation in the study. As PSA level measurement was also collected during the study and it is a commonly used indicator in prostate cancer diagnostics, I would like to include this variable as precisely as I can in a potential model that will arise in my project. Controversy that is associated with PSA level related diagnostic seems to be even more interesting when comparing to any other correlation that can be identified based on the prostate cancer dataset.
Aims
Data mining technics will be applied in order to identify any potential trends and dependencies that might exists in prostate cancer datasets. An attempt will also be made to create a model that explains observations included in datasets (decision trees, neural networks). Results of my project will be summarized in a thesis concluding my post graduate study in statistical analysis and data mining at Warsaw School of Economics, Poland.
Collaborators
individual work