ML algorithm selection flowchart

1. Define ML Problem Statement

1.1. Break down overall complex problem into fundamental problems and apply an ML solution to each fundamental problem.

1.1.1. Can the problem be solved manually or by using a conventionaly (non-ML algorithm)?

1.1.1.1. Yes

1.1.1.1.1. Apply algorithmic design patterns and best practices. Implement and test a solution in your environment without ML.

1.1.1.2. No, an ML solution is required for performance/time/complexity reasons.

1.1.1.2.1. Research and gather sufficient datasets for your problem.

2. Testing and experimenting with multiple algorithms or using ensemble learning (multiple algorithms in combination) will lead to best solution. Always experiment and seek to minimize the ML algorithm error by optimizing the algorithm hyperparameters and the the loss/cost function,based on your scenario.

3. For each of the a lgorithms presented i n red boxes, the algorithm's hyperparameters should be evaluated and optimized.