Crate vikos [−] [src]
A machine learning library for supervised regression trainings
This library wants to enable its users to write teachers independently of the model trained or the cost function that is meant to be minimized. To get started right away, you may want to have a look at the tutorial.
Design
The three most important traits are Model, Cost and Teacher.
Modules
| cost |
Implementations of |
| linear_algebra |
Defines linear algebra traits used for some model parameters |
| model |
Implementations of |
| teacher |
Learning algorithms implementing |
| training |
Holds helper functionality for Teacher algorithms |
| tutorial |
A short tutorial on how to use vikos to solve the problem of supervised machine learning: We want to predict values for a quantity (the target), and we have some data that we can base our inference on (features). We have a data set (a history), that consists of features and corresponding, true target values, so that we have a base to learn about how the target relates to the feature data. |
Traits
| Cost |
Representing a cost function whose value is supposed be minimized by the training algorithm. |
| Model |
A Model is a parameterized expert algorithm |
| Teacher |
Algorithms used to adapt Model coefficents |
Functions
| learn_history |
Teaches |