chelseaparlett / cpsc392parlettpelleriti

CPSC392ParlettPelleriti

Readings (optional, all books free)

• Python Data Science Handbook (Beginner book + code)
• Introduction to Statistical Learning (Intermediate Book + code)
• Elements of Statistical Learning (Advanced Book)

Topic	PDSH	ISLP	ESL
Pandas/Numpy	Chapters 2,3	Chapters 2.3.3
Visualization	Chapter 4 (mpl, sb)	Chapters 2.3.4 (mpl)
Linear Regression	Chapters 5.3, 5.6	Chapters 2.2.2, 3, 5	Chapters 2.3.1, 3.2, 7.2, 7.3, 7.10
LASSO/Ridge (Regularization)	Chapter 5.6	Chapter 6.2	Chapter 3.4, 3.8
Logistic Regression		Chapter 4.3	Chapter 4.4
Decision Trees/Tree Based Models	Chapter 5.8	Chapter 8.1, 8.2	Chapters 9.2, 10.10.2, 15
Naive Bayes	Chapter 5.5	Chapter 4.4.4
K-Nearest Neighbors		Chapter 2.2.3	Chapters 2.3.2, 13.3
K-Means/Gaussian Mixture Models	Chapters 5.11, 5.12	Chapters 12.4.1, 12.4.3	Chapters 8.5, 13.2.1, 13.2.3, 14.3.6, 14.3.7
DBSCAN
Hierarchical Clustering		Chapter 12.4.2	Chapter 14.3.12
Principal Component Analysis	Chapter 5.9	Chapter 6.3.1, 12.2	Chapter 14.5.1
Neural Networks		Chapter 10	Chapter 11

Resources

• Lecture Playlist
• Google Colab: a shareable online jupyter notebook. Like google docs for code (but instead of being able to see changes live, you can only see other people's changes when you save and re-open notebook). Great for collabing with other people, or for storing your code online.
• OBS Streamlabs: Great for recording your presentation.
• Slack: A place to ask me questions, meet with other students, share fun data sets...etc.
• HP Smart Apple Android: An app that will let you "scan" things to PDF using the camera on your phone/tablet. Great for sending me signed documents, or other important files.

cpsc392parlettpelleriti's Issues

Adding comparison with ideal coefficients

Insert below codes before ggplot to get the ideal parameters besides in the graph.

sim <- rbind(sim,data.frame(conames = c(NA,'A','E','I','O','U','Y','W'),
coefs = c(100,8.23,3.48,2.97,5.12,7.83,12.34,1.38),
model = 'Actual',
vowel = c(0,1,1,1,1,1,1,1)))

Insight : Linear is close, Lasso and Ridge are far below than expected. For Lasso, even the outliers are not close to approximations.

First time using Github,, thus adding the link to the code this suggestion is for.
https://github.com/cmparlettpelleriti/CPSC392ParlettPelleriti/blob/master/Extras/Linear%2C%20Ridge%2C%20and%20LASSO%20regression.ipynb