When you’ve been within the information science area for any period of time, you’ve most definitely heard this buzz time period.
The machine studying life cycle.
It sounds fancy, however that is what it actually boils right down to:
- Machine studying is an energetic and dynamic course of — it doesn’t have a strict starting or finish
- As soon as a mannequin is skilled and deployed, it would most definitely must be retrained as time goes on, thus restarting the cycle.
- There are steps inside the cycle, nonetheless, that must be adopted of their correct order and executed fastidiously
Once you Google the ML life cycle, every supply will in all probability provide you with a barely completely different variety of steps and their names.
Nonetheless, you’ll discover that for probably the most half, the cycle incorporates: downside definition, information assortment and preprocessing, characteristic engineering, mannequin choice and coaching, mannequin analysis, deployment, and monitoring.