Class Central is learner-supported. When you buy through links on our site, we may earn an affiliate commission.

PredictHealth's Multi-Factor Cost Models

Go to class Write review

Overview

Build on basic regression skills to develop more sophisticated prediction models. Learn to combine multiple factors, incorporate categorical variables, properly clean and prepare data, validate models, and engineer custom features to improve predictions.

Syllabus

Unit 1: Predicting Insurance Costs with Multiple Regression: Age and BMI Analysis

Comparing Single vs Multi-Factor Prediction Models
Extracting Meaning from Model Coefficients
Residual Analysis Reveals Model Weaknesses

Unit 2: Incorporating Categorical Variables into Insurance Cost Models

Comparing OneHot Encoding Drop Strategies
Flexible Preprocessing Pipeline Configurations
Pipeline Performance Comparison and Analysis
Pipeline Inspection and Debugging Workflow

Unit 3: Cleaning PredictHealth's Customer Database

Inspecting and Filling Missing Data
Completing the Missing Value Cleanup
Building Your Outlier Detection Function
Scaling Features with MinMax Normalization

Unit 4: PredictHealth's Model Validation Strategy: Data Splitting, Evaluation, and Visualization

Splitting Data for Model Validation
Fixing Your Broken Preprocessing Pipeline
Implementing Cross Validation for Robust Assessment
Completing Your Model Performance Visualizations

Unit 5: Creating PredictHealth's Custom Predictors: Feature Engineering for Better Insurance Cost Models

Creating Your First Age Groups
Visualizing Raw vs Engineered Features
Building Your Complete Feature Pipeline
Building Your First Predictive Model

Reviews

Start your review of PredictHealth's Multi-Factor Cost Models