lucasas/closeness-centrality

closeness-centrality

An implementation of metric called Closeness Centrality which will be used as score for a given customer.

Usage

This is a web project created using the Leiningen and Pedestal.

So the first step is start Pedestal server:

lein repl
closeness-centrality.server=> (server/start runnable-service)

This project implements a restful api. First of all, you should access the entry point / to see which routes you can follow: http://localhost:8000/

Will return:

{"_links":
  {"self":
    {"href":"http://localhost:8080/edges"}
  }
}

The endpoint /edge can be used to insert new edges. It works sending a POST request with origin and destiny:

curl -d "origin=4&destiny=2" http://localhost:8080/edges

The response will be all customers sorted by score:

[
    {
        "_links": {
            "self": {
                "href": "http://localhost:8080/customer/44"
            },
            "fraudulent": {
                "href": "http://localhost:8080/customer/44/fraudulent"
            }
        },
        "id": "44",
        "score": 0.592814371257485
    },
    {
        "_links": {
            "self": {
                "href": "http://localhost:8080/customer/88"
            },
            "fraudulent": {
                "href": "http://localhost:8080/customer/88/fraudulent"
            }
        },
        "id": "88",
        "score": 0.5857988165680473
    }
]

The fraudulent endpoint adjusts customer's score according to the following rule:

• The fraudulent customer score should be zero.

• Customers directly referred by the "fraudulent" customer should have their score halved.

• Scores of customers indirectly referred by the "fraudulent" customer should be multiplied by a coefficient F: F(k) = (1 - (1/2)^k)

  where k is the shortest path from the "fraudulent" customer to the customer in question.

Test

Just run:

lein test

Roadmap

1. Create tests for Pedestal service.clj (I've tried this but a lot of erros happened)
2. Validate origin and destiny parameters
3. Remove all memory data and use a database to store it.