We must dig deep to find insights in our data that reveal patterns and relational variables.
Advanced extraction, organization and visualizers play a major role in these discoveries.
d3VIEW currently has 16 machine learning models to utilize with even more to come.
Choose from the two supervised learning types, regression which predicts numerical values, and classification which predicts categorical values.
Or, choose the unsupervised learning type, clustering, which groups data points into clusters for a more generalized understanding of patterns in data.
This example shows the use of linear, lasso and ridge regression ML models to predict HIC values based on hood and bumper thickness.
These regression predictions illustrate their models as 3D scatter/surface plots.
Use the generated model to predict new points, as shown to the right using the HIC example from above.
Manually input single points or upload a CSV of multiple points to predict data based on the ML model.
This example shows the use of decision tree, random forest and feature importance classification ML models to predict flower class based on petal and sepal width and length.
These classification predictions illustrate their models as hierarchal data trees and horizontal bar charts.
This example shows the use of k-means and mean shift clustering ML models to predict groups based on automobile miles per gallon and weight.
These regression predictions illustrate their models as scatter plots with k-means using indicated cluster amount and means shift using indicated cluster bandwidth for grouping.
Currently, the main d3VIEW applications that can tap into these Machine Learning capabilities include HPC jobs, Simulations, Physical Tests, Databases, Workflows and Simlytiks®.
d3VIEW plans on incorporating more ML features to come which include but are not limited to:
