Algoneer is built around a handful of simple core concepts, which we introduce in the following paragraphs. These concepts are
Algorithm: An Algorithm is a well-defined method that we can train on a given DataSet to obtain a Model. An algorithm can contain parameters that modify its behavior.
Model: A Model is obtained by combining an Algorithm with specific parameters, usually by training it on a DataSet.
DataSet: A DataSet is a collection of DataPoints with a given DataSchema.
DataPoint: A DataPoint is a single element of a DataSet.
DataSchema: A DataSchema describes the format of a dataset. It can be applied to a DataSet to enforce / check the data in it, or used to generate synthetic data.
Session: A Session captures all relevant information of a given test session. It can be used to send results, parameters and data to the Algonaut API or to download information from it.
Result: A Result describes the outcome of one or several Tests performed on either an Algorithm, Model, DataSet or combination thereof.