Starting in Fall 2024, the structure of the SDS MicroMasters Program is changing. A new course will be offered—IDS.S24x Learning Time Series with Interventions—as well as two new Program Tracks, with which to complete the SDS MicroMasters Program and earn the Program Credential.
For more information about Fall 2024 SDS MicroMasters changes in general, see this FAQ article
IDS.S24x Learning Time Series with Interventions
The SDS MicroMasters Program now includes the course IDS.S24x Learning Time Series with Interventions. This course is part of the two new SDS MicroMasters Program Tracks, the Methods Track and the Time Series with Time Series and Social Science Track. It is broadly comparable to other SDS MicroMasters courses in general details.
General Information
IDS.S24x Learning Time Series with Interventions is a graduate-level course that focuses on three overlapping lines of development:
- Learning Structured Models: Focused on learning the underlying stochastic dynamic model that generates time series data, understanding how the algorithms depend on the underlying class of models adopted for the purpose, and addressing the accuracy and reliability of the models.
- Prediction: Focused on predicting the next outcome of the model processes based on past observations, making no assumptions about how the time series data was generated, analyzing Matrix and Tensor Completion Methods for providing these predictions, and analyzing the accuracy of predictions in the context of noise and missing data.
- Optimal Intervention and Reinforcement Learning (RL): Focused on understanding RL's simulator for estimating reward value for a given intervention, building on techniques from RL and the previous modules to discover how to derive new interventions & controls with better outcomes.
The Coursework will comprise three hands-on projects, in which learners will apply the knowledge gained from lectures, build models, and implement algorithms to solve problems posed—using real time series datasets.
Takeaways
You will learn how to:
- Analyze time series from the perspective of Linear Time-Invariant (LTI) systems, via methods and tools like spectral analysis
- Model time series using Autoregressive Moving Average (ARMA) and integrated processes
- Perform prediction and imputation on general time series data using matrix completion methods
- Optimize control and interventions for time series using varied dynamical programming and reinforcement learning algorithms
Prerequisites
The course assumes that learners will be prepared with pre-existing proficiency in the following academic topic areas:
- Undergraduate Python programming
- Undergraduate multi-variable calculus
- Undergraduate linear algebra
- Undergraduate probability theory and statistics
- Basic knowledge of complex numbers
These pre-requisite subjects will not be taught or reviewed within the course session, so learners not already proficient in them may expect to struggle with the course content.
Course Pacing
Like other SDS MicroMasters courses, it is Instructor-paced—Course Content will be released on a schedule unique to the course and each Coursework Assignment will have a set due date (both typically weekly). Assignment deadlines are strict, and cannot be extended. Coursework cannot be submitted for Course Score after its due date for any reason.
Who Needs to Take IDS.S24x?
As an elective Track Course, only learners pursuing one of the two new SDS MicroMasters Program Tracks—Methods or Time Series with Time Series and Social Science—need to take IDS.S24x.
IDS.S24x is not a requirement to complete either of the pre-existing Program Tracks—General or Social Sciences. Learners who are already pursuing either of these Program Tracks, and new leaners starting them, do not need to take IDS.S24x to complete the SDS MicroMasters Program and earn the MicroMasters Program Credential.
For more information about the SDS MicroMasters Program Tracks, see this FAQ article