ML Problems
An exploratory notebook collection covering data manipulation, plotting and a supervised regression example using scikit-learn.
7 commits168 source lines
Counting scope & snapshot
All commits reachable from this default-branch revision, including merges and all authors. Physical source lines counted in each physical file, including native Astro components, tests, examples and parsed notebook code cells; excluding blanks, comments, dependencies, generated files, documentation and notebook outputs.
Counted with cloc 2.10. Revision d5f240de95, observed . These are repository totals, not individually authored lines.
- Role
- To be supplied
- Technologies
- Python, Jupyter, scikit-learn, pandas, NumPy
- Impact
- To be supplied
Data and workflow
The main notebook loads scikit-learn’s California housing dataset into a pandas DataFrame, examines feature correlations, and separates features from the target. It uses an 80/20 train/test split with a fixed random seed.
Models and inspection
The notebook fits a linear regression and a random forest with 100 estimators. It contains code to calculate mean squared error and R², plot predictions against observed values, inspect linear coefficients and visualize random-forest feature importances.
A final input routine accepts housing features and produces a model prediction. This is an educational workflow, not a deployed valuation service.
Supporting notebooks
The learning directory contains NumPy, pandas and Matplotlib exercises. The declared Python dependencies also include seaborn and Jupyter.
Limits of the evidence
The notebooks were read but not executed during this portfolio update. Saved outputs are not presented as independently reproduced results. The project does not establish production model quality, external validation or a reliable basis for financial decisions.