1 – Identify the business needs and all the constraints and requirements 

(what do we want to predict? what technical environment? what timing?)

=> clear and measurable objectives must be set with the project owner

2 – Manage the availability of data and perform an analysis 

=> the data must be made easily accessible

=> their main characteristics must be understood

3 – Based on the analysis, perform the necessary transformations to make the data compatible with the requirements of the algorithms to be used 

  

4 – Select several machine learning models compatible with the data, objectives and constraints. 

  

5 – Train each model and evaluate its performance 

=> compare the performances and select the most efficient model

6 – Deploy the model and use it in production 

=> always listen to the business teams and regularly re-evaluate the model’s performance

cerveau et ampoule