Visualising Machine Learning Models
Data visualization is widely used to gain an understanding of data. Visualisation of models however, is something that is not often used. We normally limit ourselves to understanding a model based on its numerical summary. But visualising it helps us answer questions like What does the model look like?, and How will the model change when I change a certain parameter? easily.
This talk will introduce the audience to the ModelVis package in python, which has been developed to allow for convenient visualisation of regression, classification and clustering models. This package gives users the power to:
- Visualise the model in the data space
- Visualise the entire space of models
- Visualise the same model with different input data sets
The ModelVis package is a great tool for both, data scientists who wish to look deeper into the model, as well as to an end user, for whom a visual representation of the model is easily relatable.
Basic understanding of what goes into data science.
Kavya Atmakuri is a third year computer science undergrad at NITK Surathkal. She works in the fields of data science- mainly visualisation- and human computer interaction. She has a knack for art and craft.