Description
Machine Learning for Classification Problems focuses on the application and interpretation of both traditional and next-generation machine learning (also statistical learning) approaches. One of the fastest growing areas of statistics and data analysis, machine learning applications have increased rapidly within the psychological, health, and social sciences. The methods covered in this workshop focus on producing robust classifications while achieving high accuracy (versus null hypothesis testing or statistical significance. When completed, participants will have an understanding of how to choose between machine learning techniques, utilizing software such as R and Python to compare a variety of classification models.
The workshop begins with common, fundamental models that are widely used, such as logistic regression, and uses these models to introduce and understand more advanced approaches such as regularization, classification trees, and discriminant analysis. Along the way, a comprehensive approach to assessing model fit and protecting against overfitting will be developed. Finally, this workshop will also delve into the tension between “prediction” and “inference” and explores the implications for common research applications.
