Marsill/CustomerSegmentation-kmeans

Customer Segmentation Model

This repository contains a trained K-Means clustering model for customer segmentation, along with a scaler to preprocess input features.

---

Files

• kmeans_model.pkl – Trained K-Means model
• scaler.pkl – StandardScaler used to normalize features before clustering

---

Description

The model segments customers into two clusters:

Cluster	Label	Description
0	High Spender	Customers with high sales and profit
1	Occasional Buyer	Customers with lower or occasional purchases

The model was trained on the Superstore dataset.

---

Input Features

To predict the cluster of a customer, the model expects a dictionary with the following keys:

• TotalSales – Total sales of the customer
• AvgSales – Average sales per order
• TotalProfit – Total profit from the customer
• AvgProfit – Average profit per order
• TotalQuantity – Total quantity purchased
• AvgDiscount – Average discount received (0-1)

---