How ansu-john/regression-models works
Project type and purpose
• This is a data-science / machine-learning analytics project, organized as Jupyter Notebook experiments.
File structure
• Top-level items:
– Six notebooks, each implementing one regression technique.
– A resources/ folder containing CSV datasets.
• No backend/server or frontend web app – everything runs in a Notebook environment.
• External dependencies implied: Python interpreter, scikit-learn, pandas, matplotlib/seaborn.
Main components
a. Data Sources
– Ecommerce Customers.csv
– Social_Network_Ads.csv
– mushrooms.csv
b. Notebook Environment
– Jupyter server / kernel
c. Data Processing & Feature Engineering
– pandas-based loading, cleaning, transforming
d. Modeling Components (one per notebook)
– Linear Regression
– Logistic Regression
– Polynomial Regression
– Decision Tree Regression
– Random Forest Regression
– Support Vector Regression
e. Machine-Learning Framework
– scikit-learn library
f. Visualization & Evaluation
– matplotlib / seaborn plots
– metrics computation (MAE, RMSE, R²)
Relationships and data flow
• Data Processing → Modeling Components
• Modeling Components ↔ scikit-learn (calls to fit(), predict())
• Modeling Components → Visualization & Evaluation
• All notebooks run inside the Jupyter Environment
Architectural patterns / design principles
• Modular separation by technique: each notebook is a self-contained “experiment module.”
• Reusable data-loading and evaluation “sub-pipeline.”
• Clear distinction between data ingestion, model training, and result rendering.