| --- |
| license: mit |
| tags: |
| - customer-segmentation |
| - machine-learning |
| - clustering |
| - dbscan |
| - scikit-learn |
| --- |
| |
| # Customer Segmentation - DBSCAN |
|
|
| ## Overview |
|
|
| This model performs customer segmentation using the DBSCAN clustering algorithm. It identifies groups of customers based on similarities in demographic and behavioral features while also detecting customers that do not belong to any major group. |
|
|
| ## Model Type |
|
|
| DBSCAN Clustering |
|
|
| ## Features Used |
|
|
| The model was trained using the following features: |
|
|
| - Gender |
| - Ever Married |
| - Age |
| - Graduated |
| - Profession |
| - Work Experience |
| - Spending Score |
| - Family Size |
| - Age Group |
| - Family Category |
|
|
| ## Output |
|
|
| The model assigns a cluster label to each customer. |
|
|
| Example: |
|
|
| - Cluster 0 |
| - Cluster 1 |
| - Cluster 2 |
| - -1 (Noise / Outlier) |
|
|
| A label of `-1` indicates that the customer was identified as an outlier and does not belong to any major cluster. |
|
|
| ## Files Required |
|
|
| - `dbscan_model.joblib` |
| - `scaler.joblib` |
|
|
| ## Usage |
|
|
| ```python |
| import joblib |
| import numpy as np |
| scaler = joblib.load("scaler.joblib") |
| model = joblib.load("dbscan_model.joblib") |
| customer = np.array([[...]]) |
| customer_scaled = scaler.transform(customer) |
| # DBSCAN does not support direct prediction on new samples. |
| # The model is primarily intended for clustering datasets. |
| labels = model.fit_predict(customer_scaled) |
| print(labels) |
| ``` |
|
|
| ## Author |
|
|
| Mithun |
