https://github.com/devamoghs/av-cross-sell-prediction

Predict if a person who is already insured will purchase a Vehicle Insurance or not. Check Readme.md for more.
https://github.com/devamoghs/av-cross-sell-prediction

Last synced: about 1 year ago
JSON representation

Predict if a person who is already insured will purchase a Vehicle Insurance or not. Check Readme.md for more.

• Host: GitHub
• URL: https://github.com/devamoghs/av-cross-sell-prediction
• Owner: devAmoghS
• License: mit
• Created: 2020-09-21T04:03:29.000Z (over 5 years ago)
• Default Branch: master
• Last Pushed: 2020-09-21T04:16:45.000Z (over 5 years ago)
• Last Synced: 2025-02-08T16:32:31.238Z (over 1 year ago)
• Language: Jupyter Notebook
• Homepage:
• Size: 1.13 MB
• Stars: 0
• Watchers: 2
• Forks: 1
• Open Issues: 0
• Metadata Files:
  • Readme: README.md
  • License: LICENSE

Awesome Lists containing this project

README

          
# AV-Cross-Sell-Prediction

![](https://www.aureusanalytics.com/hs-fs/hubfs/Aureus%20cross-sell.jpeg?width=1254&name=Aureus%20cross-sell.jpeg)

Just like **medical insurance**, there is **vehicle insurance** where every year customer needs to pay a **premium** of certain amount to insurance provider company so that in case of unfortunate accident by the vehicle, the insurance provider company will provide a compensation **(called ‘sum assured’)** to the customer.  

# Objective

Building a model to predict **whether a customer would be interested in Vehicle Insurance** is extremely helpful for the company because it can then accordingly plan its communication strategy to reach out to those customers and optimise its business model and revenue. 

# Data Dictionay

| Variable             | Definition                                                                                                                  |

|----------------------|-----------------------------------------------------------------------------------------------------------------------------|

| id                   | Unique ID for the customer                                                                                                  |

| Gender               | Gender of the customer                                                                                                      |

| Age                  | Age of the customer                                                                                                         |

| Driving_License      | 0 : Customer does not have DL, 1 : Customer already has DL                                                                  |

| Region_Code          | Unique code for the region of the customer                                                                                  |

| Previously_Insured   | 1 : Customer already has Vehicle Insurance, 0 : Customer doesn't have Vehicle Insurance                                     |

| Vehicle_Age          | Age of the Vehicle                                                                                                          |

| Vehicle_Damage       | 1 : Customer got his/her vehicle damaged in the past. 0 : Customer didn't get his/her vehicle damaged in the past.          |

| Annual_Premium       | The amount customer needs to pay as premium in the year                                                                     |

| Policy_Sales_Channel | Anonymised Code for the channel of outreaching to the customer ie. Different Agents, Over Mail, Over Phone, In Person, etc. |

| Vintage              | Number of Days, Customer has been associated with the company                                                               |

| Response             | 1 :  Customer is interested, 0 : Customer is not interested                                                                 |

# Evaluation Metric

ROC_AUC score

# Model and Score

#### Model Type: Random Forest	

#### Public Score: 0.754661115842268	

#### Private Score: 0.758023433881523