Collaborative Science

Accelerating time to market and driving innovation with collaboration, knowledge based under-standing and prediction.

BIOVIA Pipeline Pilot Analytics and Machine Learning

BIOVIA Pipeline Pilot Analytics and Machine Learning offers a comprehensive set of learning and data modeling capabilities, statistical filters and clustering components optimized for large real-world data sets.

BIOVIA Pipeline Pilot Analytics and Machine Learning Features

  • Rapidly cluster data
  • Employ categorical learning using Bayesian statistics
  • Perform Principal Component Analysis (PCA)
  • Apply regression - Linear, partial least squares (PLS) and k-nearest neighbor (kNN)
  • Utilize Recursive Partitioning methods for building single tree or forest models. The methods can build models for single or multiple response properties.
  • Analyze variable importance to identify the most discriminating descriptors
  • Deploy model applicability domain (MAD) methods and enable cross-validation to assess prediction quality
  • Return multiple models rather than a single "best" model by creating a number of trial models. Combine multiple models into a single ensemble model
  • Employ Genetic Function Approximation (GFA) methods to perform variable selection and build multiple models, which can be combined into a consensus or ensemble model
  • Employ Pareto Optimization for multi-objective optimization problems
  • Generate interactive reports with ROC plots, enrichment plots and other visualization techniques for evaluating model quality and understanding the relationships between descriptors and responses.
  • Integrate with other statistic platforms such as R, JMP and SAS

Browse By: