kirthaniyangar.com
kirthaniyangar.com
Classification Models (Customer Churn Prediction, Cancer Likelihood Prediction, Bank Customer Analysis)
Decision Tree Analysis
Description: Comprehensive project using decision trees to classify customer behavior in telecommunications sector. Involved data preprocessing, feature selection, model optimization through cross-validation, and performance evaluation across different tree depths and sample splits.
Problem Statement: Need to develop accurate classification model for customer behavior while optimizing model parameters for best performance.
Outcome: Created optimized decision tree model achieving high accuracy in customer classification with clear feature importance identification.
KNN Implementation
Description: Development of k-Nearest Neighbors model for bank customer analysis, focusing on churn prediction. Project included testing different k values and weight strategies to optimize model performance.
Problem Statement: Required to implement and optimize KNN algorithm for customer behavior prediction while determining optimal parameter values.
Outcome: Successfully implemented KNN model with optimized parameters, providing accurate customer classification results.
Logistic Regression Analysis
Description: Advanced analysis using logistic regression to predict cancer likelihood in patients. Project compared multiple models with different variable combinations and evaluated their effectiveness using statistical measures.
Problem Statement: Need to develop accurate prediction model for medical diagnosis while identifying most significant predictive variables.
Outcome: Created effective logistic regression model with validated results and clear interpretation of odds ratios.