{"id":8185,"date":"2021-04-05T00:18:03","date_gmt":"2021-04-05T00:18:03","guid":{"rendered":"https:\/\/wealthrevelation.com\/data-science\/2021\/04\/05\/shapash-making-machine-learning-models-understandable\/"},"modified":"2021-04-05T00:18:03","modified_gmt":"2021-04-05T00:18:03","slug":"shapash-making-machine-learning-models-understandable","status":"publish","type":"post","link":"https:\/\/wealthrevelation.com\/data-science\/2021\/04\/05\/shapash-making-machine-learning-models-understandable\/","title":{"rendered":"Shapash: Making Machine Learning Models Understandable"},"content":{"rendered":"
\n <\/p>\n

By Yann Golhen<\/a>, MAIF, Lead Data Scientist<\/b>.<\/p>\n

<\/p>\n

Shapash Web App Demo<\/a><\/em><\/p>\n

Shapash<\/a>\u00a0by\u00a0MAIF<\/a>\u00a0is a Python Toolkit that facilitates the understanding of Machine Learning models to data scientists. It makes it easier to share and discuss the model interpretability with non-data specialists: business analysts, managers, and end-users.<\/p>\n

Concretely, Shapash provides easy-to-read visualizations and a\u00a0web app<\/a>. Shapash displays results with appropriate wording (preprocessing inverse\/post-processing). Shapash\u00a0<\/strong>is useful in an operational context as it enables data scientists to use explicability from exploration to production: You can easily deploy local explainability in production to complete each of your forecasts\/recommendations with a summary of the local explainability.<\/p>\n

In this post, we will present the main features of\u00a0Shapash\u00a0and how it operates. We will illustrate the implementation of the library on a concrete use case.<\/p>\n

\u00a0<\/p>\n

Elements of context<\/h3>\n

\u00a0<\/p>\n

Interpretability and explicability of models are hot topics. There are many articles, publications, and open-source contributions about it. All these contributions do not deal with the same issues and challenges.<\/p>\n

Most data scientists use these techniques for many reasons: to better understand their models, to check that they are consistent and unbiased, as well as for debugging.<\/p>\n

However, there is more to it:<\/p>\n

\n

Intelligibility matters for pedagogic purposes. Intelligible Machine Learning models can be debated with people that are not data specialists: business analysts, final users\u2026<\/em><\/p>\n<\/blockquote>\n

Concretely, there are two steps in our Data Science projects that involve non-specialists:<\/p>\n

Exploratory step & Model fitting<\/strong><\/p>\n

At this step, data scientists and business analysts discuss what is at stake and define the essential data they will integrate into the projects. It requires understanding the subject well and the main drivers of the problem we are modeling.<\/p>\n

To do this, data scientists study global explicability, features importance, and which role the model\u2019s top features play. They can also locally look at some individuals, especially outliers. A Web App is interesting at this phase because they need to look at visualizations and graphics. Discussing these results with business analysts is interesting to challenge the approach and validate the model.<\/p>\n

Deploying the model in a production environment<\/strong><\/p>\n

That\u2019s it! The model is validated, deployed, and gives predictions to the end-users. Local explicability can bring them a lot of value, only if there is a way to provide them with a good, useful, and understandable summary. It will be valuable to them for two reasons:<\/p>\n