MRC Biostatistics Unit
Events Search and Views Navigation
February 2019
BSU Seminar: “Sample size considerations for the design clinical trials – quantifying the target difference and the target no-difference”
Speaker: Prof Steven Julious, University of Sheffield Title: "Sample size considerations for the design clinical trials - quantifying the target difference and the target no-difference" Abstract: One of the most of the important steps in any clinical trial design is the calculation of the sample size calculation. From this the study timelines and financial budgets are calculated. A major driver in the sample size calculation size calculation is the target difference (for a superiority trial) or the target no-difference (for a non-inferiority…
Find out more »BSU Seminar: “Optimal Feature Selection using model-based Deep Reinforcement Learning”
Speaker: Dr Konstantina Pallas, Microsoft Title: "Optimal Feature Selection using model-based Deep Reinforcement Learning" Abstract: TBC
Find out more »March 2019
Machine learning meets statistics: Guiding medicine into the future
SOLD OUT Joint workshop, organised by The Alan Turing Institute and MRC Biostatistics Unit, University of Cambridge Traditional statistical analysis and machine learning have many things in common but usually follow paths that are quite different. For prediction, traditional statistical analysis usually begins with a theory and a model and fits the parameters of the model to the data; machine learning follows a more pragmatic approach, allowing the data more freedom to prescribe the model. Machine learning often leads to…
Find out more »April 2019
BSU Seminar: “Building Representative Matched Samples with Multi-valued Treatments in Large Observational Studies”
Speaker: Professor Jose Zubizarreta, Harvard University Title: "Building Representative Matched Samples with Multi-valued Treatments in Large Observational Studies" Abstract: In observational studies of causal effects, matching methods are widely used to approximate the ideal study that would be conducted under controlled experimentation. In this talk, I will discuss new matching methods that use tools from modern optimization to overcome five limitations of standard matching approaches. In particular, these new matching methods (i) directly obtain flexible forms of covariate balance, as specified before matching…
Find out more »