MSDS 413-DL Time Series Analysis and Forecasting.

This course covers analytical methods for time series analysis and forecasting. Specific topics include the role of forecasting in organizations, exploratory data analysis, stationary and non-stationary time series, autocorrelation and partial autocorrelation functions, univariate autoregressive integrated moving average (ARIMA) models, seasonal models, Box-Jenkins methodology, regression models with ARIMA errors, volatility models, and multivariate time series. Also included are non-linear time series models, exponential smoothing methods, random forest analysis, deep learning methods, and hidden Markov modeling. Recommended prior courses: 410-DL Data Modeling for Supervised Learning and MSDS 411-DL Unsupervised Learning Methods. Prerequisites: (1) MSDS 420-DL Database Systems or CIS 417 Database Systems Design and Implementation and (2) MSDS 422-DL Practical Machine Learning or CIS 435 Practical Data Science Using Machine Learning.

This course uses R extensively, with deep learning components in Python. Students benefit by taking the Python Learning Studio and MSDS 430 Python for Data Science prior to taking this course.

Back to main page for Analytics and Modeling.