- Pick a model
- Develop and implement algorithms that can predict outcomes based on the selected features.
- Do not think about refinement yet.
- parametric vs non-parametric models & Determining the right algorithm
- Train the model using the prepared data to learn patterns and make predictions.
Feature Importance: After training, analyze which features have the most significant impact on the model’s predictions.
Model Evaluation: Assess the model’s performance using various metrics to ensure it meets the desired accuracy and reliability.
- Choose the best-performing model based on evaluation metrics and optimization results.
- Cross Validation: Evaluate the model more robustly by splitting the training data into smaller chunks and training the model multiple times.
Model Refinement:
- Model Interpretability: Utilize tools to understand and interpret the model’s predictions, ensuring transparency and trustworthiness.
- Model Ensembling: Combining models.
- Deploy the model into a production environment where it can be used.
Model Observability: Monitor the model’s performance over time.
Model Retraining: Retrain on new data.