Date: October 6
More on the Bootstrap:
Bootstrap, originating from the statistics field, refers to a method used to estimate the distribution of a statistic (like the mean or variance) by resampling with replacement from the data. It allows the estimation of the sampling distribution of almost any statistic. The primary advantage of Bootstrap is its ability to make inferences about complex statistical measures without making strong parametric assumptions.
- Resampling with replacement: This means that in a dataset of ‘n’ values, every time a sample of ‘n’ values is drawn, any particular value might be selected multiple times.
- Non-parametric Bootstrap: This involves straightforward resampling.
- Parametric Bootstrap: Assumes data comes from a known distribution and estimates the parameters.
- Smoothed Bootstrap: Adds random noise to the resamples.
Discussed Project 1 Doubts:
During our discussion on Project 1, several uncertainties were clarified:
- Scope & Requirements: We revisited the primary objectives of the project, ensuring all participants understood the expected deliverables and performance criteria.
- Dataset Concerns: Some doubts were raised about data integrity, missing values, and the potential need for data transformation or normalization.
- Implementation Details: Questions regarding certain algorithms, tools, and libraries to be used were addressed. We discussed possible pitfalls and alternative approaches if our primary strategies do not yield the desired results.
- Timeline & Milestones: We reiterated the importance of adhering to the project timeline, ensuring that key milestones are met on schedule. Concerns related to resource allocation and task delegation were also addressed.