It is difficult to find the right formula for recommendations. For this, an evolutionary algorithm is used.
Table 8.1. Evolution config
| Attribute | Description | Default recommender | Default ml |
|---|---|---|---|
| generations | How many generation (iterations) we will have. | 100 | 10 |
| select | The size of the population after each selection | 16 | 16 |
| elite | The number of best fitted that will not mutate | 1 | 1 |
| elitecloneandmutate | The number of elite that will be cloned mutated (unofficial) | 1 | 1 |
| crossover | The number of crossover to make a new individual | 2 | 2 |
| mutate | The number to mutate | 2 | 2 |
| generationcreate | The number to create per generation (unofficial) | 1 | 1 |
| useoldelite | Use old elite if exists (unofficial) | true | true |