- Instructor: William B. Rossow, Distinguished Professor of Electrical Engineering
- Course: Graduate Seminar (EE G6904): Advanced Statistics and Non-Linear Analysis. (Fall 2015)
- Class Hours: Wednesdays 2:00-4:45
- Room: Steinman 512
- Office Hours: Wednesdays 12:00-2:00
- Email: wbrossow@ccny.cuny.edu
- Syllabus
- Satellite remote sensing datasets are usually very large in volume and involve measurements that have complex, multi-variate and non-linear relationships with the physical quantities to be obtained from them. Analysis methods that are commonly in use today employ almost exclusively linear, uni- or bi-variate statistical approaches that are a logical first approach but can no longer advance Earth studies any farther. This seminar will, after briefly reviewing standard (linear) statistical analysis methods, focus on the use of neural networks as a statistical analysis tool. Use of neural networks will also serve to illustrate a whole class of related analysis tools. Class time will be occupied by some lectures and student presentations of the results of working through real examples based on satellite observations of Earth.
- Learning Goals: To understand the conceptual foundations of statistical analysis, especially their similarity across various methods, to understand how to apply various analysis methods in real-world situations where data quality can be an issue and to convey the understanding that these analysis methods, alone, do not provide a scientifically meaningful explanation, only a description of the data.
- Academic Integrity
- Grades: Based on combination of Homework, Attendance, Participation in Class Discussions and Final Presentation.Late Homework will be reduced in grade by one step (e.g. from A to A-) for every week of delay.
|
