ITHMAAR AI

OVERVIEW

While providing all of the basic functionality the Islamic finance industry requires, Ithmaar is also one of the few companies harnessing Artificial Intelligence as a means of predicting risk and informing strategic decisions.

OVERVIEW

Future is here

Ithmaar delivers innovation and effectiveness, both in the MENA region and worldwide. We offer next generation solutions within the market of Islamic Finance. Our team stays on top of the new development technologies and trends of the industry and employs them to build fully Sharia compliant products.

While providing all of the basic functionality the Islamic finance industry requires, Ithmaar is also one of the few companies harnessing Artificial Intelligence as a means of predicting risk and informing strategic decisions. Powerful analytics and extensive statistical data tools make it easier than ever to monitor the key metrics you specify. With such potent and comprehensive analysis at your fingertips, you will be well positioned to take advantage of future opportunities others may miss.

Islamic Finance is poised to take its rightful place among the world’s best financial systems, and our mission is to provide the technology to optimize the experience of every organization and individual within the sector without sacrificing convenience or custom.

The world of Fintech
is always on the move

What is Ithmaar AI?

Ithmaar Machine Learning Model

Ithmaar has succeeded in developing a machine learning model to predict the extent to which individual customers pay their installments at due date.
The model determines customer payment behavior from existing data to predict that of new customers. The input for learning is a large set of samples, each sample containing certain customer attributes along with a flag 1 to denote pay-on-time and 0 to indicate otherwise.
Customer attributes may vary from one organization to another, but typically the set includes nationality, age, education level, communication language, etc. The organization is responsible for deciding whether an existing customer is paying on time or not. For example, in the case of 60% of the installments being paid on or before the due date, the customer here is considered to have paid on time.
How it works?

Proof of Concept Steps

To start a workshop, an organization is required to prepare both a training dataset and a test dataset, each containing the same set of attributes for the customer and a result flag as mentioned above. It is not required to include customer code, name or any other identifying attributes. The reference for dealing with samples is just the sample number. Standardly, training dataset contains 70K samples while the testing dataset contains 30K samples. The larger the dataset size, the more accurate the prediction.
An important feature that should be present in the training dataset is for it to be balanced, i.e., the number of customers who pay on time is almost equal to that customers who do not. This feature is of major significance to reach a more accurate prediction. An additional important feature is that most customers should ascribe a value for each attribute, e.g., all customers should have a nationality, age, etc. Similarly, the more data gathered regarding the attributes the more accurate the prediction.

Take advantage of future opportunities others may miss

Workshop steps

  1. The organization provides the training dataset (attributes and results)
  1. The organization provides the testing dataset (attributes only).
  1. dddd
  1. Ithmaar model predicts the result for each sample in the testing dataset.
  1. The organization calculates the accuracy of predictions in accordance with an accuracy measure: (true positive + true negative) / number of samples.
  1. The organization provides the testing dataset (attributes only).
Download Brochure
Request A Demo Now