UofT-DSI | LCR - Assignment 3

[amera6]

What changes are you trying to make? (e.g. Adding or removing code, refactoring existing code, adding reports)

I inputted my own code and written answers in the designated sections of the assignment_3.ipnyb notebook. Specifically, I applied k-means clustering on the wine dataset from Scikit Learn to identify patterns among the features and applied bootstrapping on the same dataset to estimate the mean color intensity of the wine. I also answered questions in prose regarding standardization, pattern recognition, ideal cluster numbers, confidence intervals, and estimate stability/reliability.

What did you learn from the changes you have made?

I learned how to run an algorithm to visualize pair-wise associations for features in a dataset, as well as how to apply a clustering and bootstrapping algorithm in Python. Most of my statistical background is in R so it was neat seeing the parallels in how these techniques are coded.

Was there another approach you were thinking about making? If so, what approach(es) were you thinking of?

There was no other approach that I was thinking of using for answering these questions, as these were relatively straightforward prompts.

Were there any challenges? If so, what issue(s) did you face? How did you overcome it?

I did not encounter any challenges apart from confusion as to why the set.seed() function was placed after defining the variable for the bootstrap means in the original assignment notebook. I was also confused as to whether the instructors wanted an array of bootstrap means or if they wanted a point estimate that would be the average of all bootstrap means. I just assumed the latter option given that an f-string was used.

How were these changes tested?

I re-ran the cells after adding an additional line of code, and iteratively revised the code as needed.

A reference to a related issue in your repository (if applicable)

Checklist

• I can confirm that my changes are working as intended

[anjali-deshpande-hub]

Well done! Everything looks good except one small thing.

I saw that you chose a random sample of 40 and then bootstrapped the subset instead of using the entire dataset, which wasn't required. Bootstrapping a random sub-sample of the population inflates uncertainty because your initial 40 may not fully represent the whole dataset.