Class Description
The workshop will focus on knowledge extraction and discovery from data, using statistical tools and machine learning algorithms. The students will be required to design and implement such systems and present their results in class.
Meeting Schedule
| # |
Date |
Class Details |
Lecturer |
Files |
| 1 |
30/10/2016 |
Introduction: Intro to data science, project details, important Dates |
Daniel Deutch |
Slides |
| 2 |
06/11/2016 |
Hands-On Data Science in Python : iPython,Jupyter Notebook, Numpy, Scipy, Scikit-Learn, Pandas |
Amit Somech |
Slides
Material (Notebook, data files)
|
| 3 |
11/12/2016 |
Preliminary Demo: Demo presentations by students, describing the outlines of the project (Problem formulation, tools and techniques, etc.) |
|
|
| 4 |
01/01/2017 |
Status Meeting: Teams will report their current status of the project |
|
|
| 5 |
26/01/2017 |
Final Project presentations (Thursday, 3-6pm): Final presentations: Problem,model,techniques,current results |
|
|
| 6 |
13/03/2017 |
Projects Submission Deadline |
|
|
Notifications
| Date |
Notification |
| 25/10/2016 |
Final project guidelines: here (PDF) |
| 25/10/2016 |
Final project grading sheet: here (PDF) |
| 30/10/2016 |
First task: (1) Send us details of your group members by 06/11/2016
(2) Browse through the data, and focus on a preliminary goal, directions and tools.
(3) Send us 3-5 slides describing (2) by 20/11/2016
(4) Ask to meet us if you're stuck.
|
| 30/11/2016 |
Preliminary demo guidelines:
- Presentations will be in front of class, schdule will be published on December 8th
- Each presentation is 10 minutes
- Should include the following:
- The goal of your project, and why is it interesting/important
- The dataset segements you chose to deal with, and a short explanations of what they contain
- A forumlation of a machine-learning problem (e.g "predict future GDP of first-world countries")
- Problems you have encoutered so far (e.g missing data) and how do you intend to deal with them
- An outline of your implementation plan
|
Resources