Please Wait

Platform X About Us Our Partner Careers Newsroom Resources Contact Us
cancel

Request for a Demo

Understand more about Platform X

What is Platform X - AutoML?

AutoML is an Automated Machine Learning solution to empower domain and business experts in banks to use Machine Learning (ML) and Artificial Intelligence (AI) for creating business values. Business and analytics teams can develop and implement advanced AI projects without coding and understanding the underlying algorithms.

Automated Model
Tuning and selection

CreditNext automatically performs hyperparameter tuning and finds the top N models on a given dataset.

The solution also supports manual tuning for expert data scientists. 

It also recommends the best model from N candidate models based on performance on the out-of-sample dataset.

Structured Machine
Learning Pipeline

Adopt best practices to build ML models to improve model generalization to the out-of-sample dataset and avoid overfitting.

Model Output
Explainability

Easily explain the output of a model, to ensure full transparency for regulatory and internal purposes.

Explore a model, its important features and easily understand why it made a specific prediction using multiple visualizations.

Model
Documentation

Customizable model documentation template to generate detailed documentation for a selected model.

Integrated
MLOps

Automatically logs and tracks model hyperparameters, metrics and artifacts to easily replicate model output at a later stage.

Challengers
Models for
Traditional Method

Quickly generate benchmark challenger models for your existing scorecards and models to identify areas for model performance improvement. 

Transparently identifies important features that can be incorporated to existing models and build challenger models in a fraction of the time it takes to build models using traditional workflow.

EDA, Visualizations
and Reports

Pre-configured reports and visualizations used in the financial industry to speed up EDA, feature transformation, and feature selection.

Model
Deployment

Automatically deploy the model as API service or as the model object with both on-premises and on-cloud deployment options.

Capabilities of Platform X AutoML

Decide on the optimal provisions number from a large number of alternatives.

Data
Requirement

Data sample of few hundred data points is sufficient for building ML models for most business problems.

Algorithm
Coverage

Supports traditional algorithms such as Linear Regression and Logistic Regression as well as advanced algorithms such as ElasticNet, Gradient Boosting, XGBoost, Random Forest, Deep Learning

Use
Cases

Easily build high-performance scorecards and models for underwriting, risk scoring, collections, marketing analytics, channel analytics, customer analytics, etc.

Technical Features

Platform X makes use of open source software and is based on special commodity hardware. Platform X ECL computation engine is an Advanced Dynamic DAG compute engine that runs on Apache Spark which allows multiple methodologies and iterations on the fly. It has a browser based easy to use front end UI that complies with security and audit requirements.

  • Multiple teams can work together- easing the workload and support for ownership of results.
  • Platform Framework creates new computation sequences, models and result data sets without coding.
  • Speed of execution allows what-if analysis and rapid exploration of alternatives.
  • Metadata allows for rapid changes to methodology and approach when required by the regulator or business.
  • Explains the variances and results/details to management and regulators without any additional manual work.

Brochures