Reference no: EM133480740
Problem: Using Support Vector Machine (SVM) method in Python
The goal of this project is to determine which risk factors are the most significant predictors of a patient developing coronary heart disease (CHD) within 10 years.
1) How might someone use your model to make a prediction in practice
2) How did you split your sample?
Estimate/Test samples
Did you stratify the sampling?
Upsampling or Downsampling?
Other splits?
I. What variables are expected to make an impact in your model (just a few)
1) This could be from data visualizations / stepwise investigations / decision trees (
II. Example: We expect studying, tutoring, completing assignments.... Will lead to higher grades.This was determined from visuals, domain expertise, etc.
1) You should include a key visual or two from your presentation in these slides.
2) Any data transformations/standardizations to match assumptions?
3) What variables did you try and were removed? Why? (this is more for the presenter)
4) Model iterations? What did you try? (this is more for the presenter, k, act functions)
5) How accurate and by which metrics?
III. At this stage which model is best? How do you know?
1) Do you think this is an adequate solution to your problem? Can you do better?
2) Any challenges?
3) How might you improve?
4) At this stage which model is best? How do you know?
IV. Do you think this is an adequate solution to your problem? Can you do better?
1) Any challenges?
2) How might you improve?
3) Ideas for next steps
4) How could your initial model be improved (could be a sub-question)
5) Additional models you hope to try (new methods)
6) Division of data to create more targeted predictions (how will you split the data)?
7) Adding/removing data/sensitivity?
8) Testing on new data sets?