This is repository for basic interaction with Amazon Fraud Detector

Prerequisites

Create an AWS account

1. Open https://aws.amazon.com and then choose Create an AWS Account.
2. Follow the on-screen instructions to complete the account creation. Note your 12-digit AWS account number.

Create an IAM user and assign required permissions

To use Amazon Fraud Detector, you have to set up permissions that allow access to the Amazon Fraud Detector console and API operations. You also have to allow Amazon Fraud Detector to perform tasks on your behalf and to access resources that you own. The following describes how to create an IAM user and assign the needed permissions.

1. Open the IAM console.
2. In the navigation panel, choose Users and then choose Add user.
3. For User name, enter AmazonFraudDetectorUser.
4. Select the AWS Management Console access and Programmatic access check boxes, and then configure the user’s password.
5. (Optional) By default, AWS requires the new user to create a new password when first signing in. You can clear the check box next to User must create a new password at next sign-in to allow the new user to reset their password after they sign in.
6. Choose Next: Permissions.
7. Choose Create group.
8. For Group name enter AmazonFraudDetectorGroup.
9. In the policy list, select the check box for AmazonFraudDetectorFullAccessPolicy and AmazonS3FullAccess. Choose Create group.
10. In the list of groups, select the check box for your new group. Choose Refresh if necessary to see the group in the list.
11. Choose Next: Tags.
12. (Optional) Add metadata to the user by attaching tags as key-value pairs. For more information about using tags in IAM, see Tagging IAM Users and Roles.
13. Choose Next: Review to see the User details and Permissions summary for the new user. When you are ready to proceed, choose Create user.
14. After this you will see credentials for this user - better to save it. You will need it further.

Get and upload example training data

1. Download this file, unzip and use registration_data_20K_minimum.csv.
2. Create an Amazon S3 bucket:
   • open the Amazon S3 console;
   • choose Create bucket, and perform the steps to create your bucket. You must choose an AWS region where Amazon Fraud Detector is currently available: US East (N. Virginia), US East (Ohio), US West (Oregon), Europe (Ireland), Asia Pacific (Singapore) or Asia Pacific (Sydney).
3. Upload a training data file (from 1 step) to your Amazon S3 bucket.
4. Remember or note S3 URI for this uploaded file.

Environment creation

1. First of all, download and install python for you OS.

2. Next step is installing virtual environment library.

pip install virtualenv

3. Create and activate virtual environment:

python -m venv path-to-venv-folder
source path-to-venv-folder/bin/activate

4. Install all requirements

pip install -r requirements.txt

5. Create .env file

sudo mv .env_template .env

6. Fill .env file with right data.

Use Amazon Fraud Detector

This step will create all components that Amazon Fraud Detector use, fraud detection model and detector for its own.

python main.py

Useful notes about data

The training dataset must contain the following headers:

• EVENT_TIMESTAMP: вefines when the event occurred. For more information, see Event Timestamps Format.
• EVENT_LABEL: сlassifies the event as fraudulent or legitimate. The values in the column must correspond to the values defined in the event type.

Also you need to understand connection between you code and data. In 17-32 lines of the main.py script you can see code for variables. name that you defined should be same as columns in your training data.

More about data preparation you could find here.

Testing

This step test your fraud detection model on one sample.

python test.py