session_​02_​6th_​Jan_​2025_​linear_​regression

<p>This workflow demonstrates a fundamental end-to-end Machine Learning pipeline using <strong>Linear Regression</strong> to predict retail sales. It covers data ingestion, feature engineering via aggregation, data blending, and model training. The project is designed to help users understand how store-level characteristics (like average price and historical sales) influence individual transaction performance.</p><p>🧩 Workflow Steps</p><ol><li><p><strong>Data Loading</strong>: Reads retail transaction data (including price, sales, and brand) from a CSV source.</p></li><li><p><strong>Feature Engineering</strong>: Uses the GroupBy node to calculate store-specific metrics ($Mean(sales)$ and $Mean(price)$). This adds "contextual" features to the model.</p></li><li><p><strong>Data Integration</strong>: Employs a Joiner node to merge the original transactional data with the newly calculated store-level averages based on the store ID.</p></li><li><p><strong>Model Training</strong>: The Linear Regression Learner is configured to predict sales. It automatically handles categorical variables (like brand) and provides a detailed statistical summary of coefficients, p-values, and R-Squared.</p></li></ol><p>📊 Key Features Used</p><ul><li><p><strong>Low-Code ML</strong>: No manual encoding or scaling required; KNIME handles the heavy lifting.</p></li><li><p><strong>Feature Enrichment</strong>: Blending raw data with aggregated statistics to improve predictive power.</p></li><li><p><strong>Statistical Transparency</strong>: Easy access to the "Linear Regression Result View" to interpret the impact of promotions and pricing.</p></li></ul><p>🚀 How to Use</p><ul><li><p><strong>Data</strong>: Ensure the ojclientsales.csv is available in the configured path or replace it with your own retail dataset.</p></li><li><p><strong>Execution</strong>: Run the workflow up to the Linear Regression Learner node.</p></li><li><p><strong>Analysis</strong>: Right-click the Learner node and select <strong>"View: Linear Regression Result View"</strong> to analyze the coefficients and model significance.</p></li></ul>