BUEX-C531 Predictive Analytics/Data Mining
- 12 weeks
- 3.00 credits
This course on data mining and predictive analytics provides students with both the conceptual underpinnings of a broad variety of data mining models as well as experience with analyzing real data sets.
Topics include: data preparation, cleaning, and exploratory analysis using data visualization and descriptive statistics; applications of multiple regression for numeric prediction; building predictive models using logistic regression, k-nearest neighbors, Naive Bayes, classification and regression trees, neural nets, discriminant analysis, advanced predictive techniques based on ensembles of predictions such as bagging and boosting, and selected time series forecasting methods; finding patterns in data using unsupervised models including principle components and cluster analysis; evaluating the performance of predictive models using training, validation, and testing data subsets as well as k-fold cross-validation; evaluating the performance of predictive and classification models using Receiver Operating Characteristic (ROC), lift charts and statistics on confusion matrices.
The class sessions provide overview of the theory behind each model as well as demonstrations using Excel, Jupyter, and other tools as needed.