Abstract:
|
Statistical learning and data mining techniques are applicable to a diffuse range of problems in which inference must be performed in the presence of uncertainty. Successful techniques include traditional Bayesian or frequentist probability models, as well as nonprobabilistic methods such as kernel machines and many of the decision tree methods. The success of a particular technique may depend on the subject domain in which it is applied. For this reason, it is important for data scientists to be aware of the variety of models and algorithms available, the equivalence relationships between these models, and the domains in which they have been successful. This roundtable will bring together attendees interested in discussing their current practices in data science and any innovations required to make a technique excel for a particular problem. Crucially, attendees are encouraged to share the titles of any reference material they found helpful in getting a new user of the technique up to speed. From this, attendees can expect both a lively discussion, along with the first steps of the subsequent learning process.
|