Introduction
LunchBoxML’s components utilize Accord.NET and Microsoft ML.NET.
The documentation for these components is organized by each of the grasshopper submenus.
Jump to a section
Accord.NET Components
Refer to Accord.NET’s docmentation for detailed information.
This set of components is subdivided into four areas by topic.
Regression Analysis is a statistical technique that is useful for exploring the relationships between two or more variables in a dataset. These components each accept similar basic inputs.
The Result (output) is a prediction for the Test Data (input) based on the relationships in the training data provided.
Clustering and Classification Algorithms are useful for assigning data into groupings.
The third subsection contains utilities related to structuring and data sets as well as training algorithms.
The Neural Network components allow users to test and train Neural Networks.
The Trainer uses a Learning Algorithm to fine tune the network to achieve an output that is aligned with the training data set. There are several Learning Algorithms available, which can be selected by right-clicking on the input field.
Similarly, the Trainer allows users to select an Activation Function from a list of options. Users may find that experimenting with these selections (in addition to the other input options) in a trial-and-error fashion will help to discover a combination that produces desirable results.
The Tester component will run the Neural Network algorithm. Users can pass the “Trained Neural Network” output field from the Neural Network Trainer component into the Tester. The input for “Test Input Data” should be similar to the Training Inputs used for the trainer.
Data Sets Components
These components support data management functions within grasshopper.
Refer to ML.NET documentation for detailed information
Models Components
Refer to ML.NET documentation for detailed information
Trainer Types Components
Refer to ML.NET documentation for detailed information
