Predicting Hospital Readmission of Diabetic Patients Details

Contents

• Dataset Description
• Dataset Preparation

Transform data type, Deal with the missing value, Recoding and collapsing features, Categorization, Remove the outliers

• Feature Selection

Boruta algorithm *Analytical Teachiques Split the dataset into training data and test data, Data balancing, k-Fold Cross-validation

• Models and evaluation
• Model Comparison *Conclusions

Medical insurance cost with EDA + OLS Regression Details

• Ordinary Least Squares (OLS) Regression: Simple Linear Regression, Polynomial Regression, and Multiple Linear Regression
  
• Exploring and Preparing the Data, Model Building, Variable Importance, Regression Assumptions, Improving Model performance
  
• Log Transformation, Outliers Remove Function, Multicollinearity Check, Statistical Interpretation

Predicting Customer Risk by Home Equity Loan Details

• Data Pre-processing
• Imbalance Data (Decision Tree)

Undersampling, Oversampling, Both, SMOTE
Performance Metrics: Accuracy, Error Rate, Specificity, Precision, Recall(Sensitivity), F Measure, ROC, AUC

• Build Model (Logistic Regression)
• Model Diagnostics

VIF, Cutoff, Misclassification Error, Confusion Matrix, Concordance

Health Insurance Data -- EDA/Managing (ggplot2) Details

• Visually checking distributions for a single variable
• Visually checking relationships between two variables
• Plotting data with a rug
• Cleaning data
• Data Transformations

Analyzing Biological Streams using Multiple Linear Regression

• Check the Corralation Cofficient (visualization) paris.panels(), chartCorrelation(), ggpairs()
  
• Logarithmic Transformation, Variable Importance, Regression Assumptions(Residuals, Cook's D)

Breast Cancer Wisconsin - Classification

• Logistic Regression, Decision Trees, Conditional Inference Trees, Random Forest, Support vector machines, Choosing a best predictive solution