DM-SuiteAI Application API ARTIFICIAL NEURAL NETWORKS BIG DATA BUSINESS INTELLIGENCE COMPUTER VISION DATA ANALYSIS DATA MANAGEMENT DATA PREPARATION DATA QUALITY DATA STRATEGY GRAPH HEALTH IOT KNOWLEDGE BASE MODEL PERFORMANCE MANAGEMENT MODEL VALIDATION PROBABILISTIC MODEL PROCESS OPTIMIZATION SMART CAMERA SOFTWARE SYNTHETIC DATA VIDEO ANALYTICS
dm-suite is a framework for analyzing the predictive potential of generic feature sets for a wide variety of use cases
The basic idea is to make the process of describing an item, understood as a unit of analysis, as flexible and generic as possible, allowing the user - through a plugin mechanism - to easily develop or adapt his or her own methods of data acquisition and feature extraction.
For each specific analysis project, it will be able to configure how a single item is represented and to create a universe of items with similar descriptions. Different items can be associated with one or more types of classification by labeling and divided into different training and validation datasets.
The common descriptive base will then make it possible to apply univariate and multivariate analysis techniques, to analyze which of the features describing an item are the most efficient in predicting a given classification, to develop different predictive models accordingly, and to analyze their potential by calculating different validation metrics.