Presentation Date:
Isaac Kohane presents “Interesting early results in data science that require new statistical methods”
Title: “Interesting early results in data science that require new statistical methods”
Abstract: In the brave new world of biomedical data science, new sources of data emerge seemingly every year from Twitter to genomes to weather to drug habits and/or doctor preferences. I will outline several interesting and apparently impactful findings that have emerged as a result of analyses of these data, both individually and jointly. I will then follow the discussion of these early successes with an outline of significant unanswered methodological challenges requiring a systematic and sound response if further progress is to be achieved. In particular, I will focus on those challenges which I believe are of the most interest to the biostatistical community.