Quickstart

This page introduces the basic concepts related to data analysis and processing for clinical databases. The API has been originally developed for various machine learning tasks in a context of early drug development for cancer, but it is intended to be flexible and adaptable for other similar problematics.

The objective of the API is to make available the tools needed to solve the issues raised by the complexity of Electronic Health Records (EHR), that will be introduced in intro .

Introduction

The usage of EHR has been widely adopted around the earth, which has led doctors and statisticians to mine them to improve the care of patients. However, medical health records are very difficult to tackle since they contain all the difficulties that exist in data analysis: * sparcity * high cardinality categorical features * unstructured data (text, images) * temporality of the events

All those issues make it hard to initiate a machine learning project using clinical data, for those reason, we aim at providing the right tools to preprocess such databases with efficiency and simplicity.