This project applies machine learning and linear algebra to real-world problems using data from the Sure Tomorrow insurance company. Key tasks include customer segmentation using k-Nearest Neighbors (kNN), benefit prediction with kNN classifiers and linear regression, and data obfuscation with matrix multiplication to ensure privacy without compromising model performance. Highlights include an F1 score of 0.95 for kNN on scaled data and successful obfuscation maintaining identical RMSE and R² values.
👩🏽💻 Linear Regression Model ✖️ Matrices and Matrix Operations ➡️ Vectors and Vector Operations 🏙️ Manhattan Distance 🔴 Dot Product 📐 Euclidean Distances 🏡 K Nearest Neighbors Algorithm
- This project uses math, numpy, pandas, seaborn, display, LinearRegression, f1_score, mean_squared_error, r2_score, train_test_split, NearestNeighbors, KNeighborsClassifier, and MaxAbsScaler. It requires python 3.11.