How to use GiniMachine
Step 1
Collect & Prepare Data

Use historical data with columns containing attributes and rows containing records.

Each record need to have a clear status known
in advance, such as 1/0, good/bad, repaid/overdue.

No personal data is required: if it occasionally appears in the dataset, the system can
automatically identify personal data and ignore it.

A recommended minimum for a quality model is 1000 records.

Supported data formats: text, numbers, dates.

The system is trained to process missing fields.

Supported file formats: .xls/.csv.

Here is an example of a dataset for a lending business

Training records should include data known at the time of the application process.
For example, the dataset should not contain the date of the last loan payment from the training record.

Each application must have separate records. For example, if the customer has several
applications, several relevant records need to be created.

What are the recommended attributes?
  • Social and demographic data about applicants: gender, age, education, marital status, number of children.
  • Geographical data: residence, country, city.
  • It’s recommended to include no more than 50 attribute values. For example, if all borrowers are from metropolitan areas, it’s better to include only a city district and omit street addresses.
  • Employment data: profession, present employment duration, total work experience.
  • Data about employers: industry, company size, location: city, region.
  • Credit history data: the total amount of debt, number of open contracts, a total number of delinquencies.
  • Parameters calculated by the lender: debt-to-income ratio, total loan debt.
  • Alternative data: behavioral data, payment data from telecom and utility services providers.
Step 2
Build and Deploy a Model
As an ML-powered no-code solution, GiniMachine analyzes data automatically and builds a model in a few minutes completely on its own.
  1. Upload the file or paste the URL
  2. Build a new model
  3. Get a detailed report with user-friendly explanations to evaluate the model.

When the model is built, reviewed, and validated, it is ready for calculations and real-time predictions, and can instantly be deployed and integrated into your scoring process.

GiniMachine users can build an unlimited number of models at no additional cost.
Also, they can import and export models between GiniMachines deployed on different servers or for different business lines if they need to.

Step 3
Evaluate a model

After the model is built, GiniMachine provides insights and diagrams. They demonstrate the
predictive power of the model, the importance of certain attributes in your dataset, the<
recommended cutoff value, profit forecast, etc

About the Gini Index

The Gini Index shows the predictive power of the model and its usefulness in predicting results.

When it is higher than 0.6, it means the model is good enough to have a commercial implementation.

If the index is lower than 0.6, you may need to add more historical data, experiment with the
number of columns (add more or remove unnecessary ones), or make the dataset more
heterogenous if there are too many records with the same outcome.

Step 4

Your model is almost ready for scoring.

Before the start, you need to match the scoring numbers with ‘yes’ or ‘no’ decisions based on your
business objectives. The cutoff value slider will help you with that.

Select the cutoff value

1. Move the cutoff value slider to reach the balanced state<

2. Check out predictions and economic effects changing in the real-time<

3. Get ready for building as many powerful scoring models and you need

By playing with the cutoff value, GiniMachine users can completely change the confusion matrix
searching for the most profitable compromises between increased risks and missed revenues.

A balanced cutoff value means it meets your business strategy, priorities, and risk tolerance.
A neat and convenient GiniMachine interface allows experimenting in real-time to develop an optimum approach to scoring.
Start scoring

To start scoring you need to upload the data file or paste the URL. The frictionless flow and intuitive interface of the application make this process quick and extremely easy.

When the applications are scored, users can check the scoring details on each applicant and see which attributes had the most impact on their score:

Step 5

At some point, predictive models may steadily lose their predictive power and need to be updated. It may happen due to the data drift or due to the concept drift.

The data drift takes place when the input data has changed, and the distribution of the variables is meaningfully different. For example, the business started working with borrowers from a different region, so a new column appeared in the dataset. However, the model will still perform well on the data similar to the “old” one.

The concept drift occurs when the model’s patterns learned are no longer relevant. In contrast to the data drift, the applicant’s data structure may even remain the same, but the relationships between the model inputs and outputs change. For example, economic crisis

Step 6
Update your models

How to find out the right time for the model update?

Check out the dynamics and the rejection rate in the Monitoring tab of the GiniMachine interface. In case of sudden pattern changes, some investigation and model updates may be required.
To sum it all up

GiniMachine is a no-code AI-based platform that can be used for scoring, predictions, and decision-making.

The system uses historical data to build high-accuracy predictive models in minutes.

It is ready-to-use and has an intuitive interface with explanations. GiniMachine brings the power of AI/ML to serve your business and, at the same time, it requires no data scientists or machine learning engineers on board to operate it.

Got a question?
Contact us


First & Last Name

Company Name

Email Address