FAQs

The FutureAnalytica (FA) architecture is based on a modular approach, where various modules communicate inside the FA ecosystem via API calls. The additional module can be added or customized easily. These modules run over the big data pillars in order to scale horizontally.

The FA platform has various high-end machine learning (ML) algorithms – supervised, unsupervised and data engineering. In the AutoML mode, the system runs various models & tweaks in order to maximize the objective – e.g. accuracy, F1, etc.

The FA platform allows you to quickly build AI/ML models using AutoML technology. You can further customize & tune the AI/ML models without writing any code. The end-to-end data science journey is supported by the platform in a fully–/semi–customized manner, along with support from business analysts and experienced data scientists. With a unique one-click integration, you can share real–time insights with your decision-makers and drive the AI revolution in your organization in just days, not years.

The FA platform allows the users to connect various data tables from the Data Lake to automatically build features to enrich the data further for machine learning. The generated features are then selected, and the final features are presented to the end users with a detailed feature report.

FA’s platform supports various data files (local & remote locations), data sources, and databases. The data source can be connected in batches and in a stream mode.

You need at least 35 or so data points for regression and around 100 data points for classification problems, to allow FA’s AI platform to help you derive/produce something useful from your data. 

The FA platform automatically handles the derived dataset during the development and production phase.

The FA platform enables the quick development of AI models, using high-end AI/ML algorithms, along with a detailed & transparent model result. The platform is meant to accelerate the process of model building and deployment for enterprises. FA does not control the model usage, and won’t be liable if any of your AI models fail.

As of today, one can write SQL scripts to run data engineering on datasets.

The FA platform is technologically agnostic, as far as hosting the platform is concerned. It can be hosted with any of the cloud providers, as well as on-premises for a client.

The FA platform’s AutoML handles any data imbalances automatically. Advanced users can apply their own data imbalance solution(s) in the data lakehouse using scripts or using a menu-based data prep.

The AutoML model can be tailormade to fit your needs perfectly. It can be further fine-tuned using the Grid search.

No, not directly. However, you can leverage specific APIs to connect to various services.