Question

Random search differs from grid search in that you no longer provide a discrete set of values to explore for each hyperparameter; instead, you provide a statistical distribution for each hyperparameter from which values are randomly sampled and set the total number of combinations you want to try. Bayesian techniques differ from random or grid search in that they use past evaluation results to choose the next values to evaluate. The idea is to limit the number of expensive optimizations of the objective function by choosing the next hyperparameter values based on those that have done well in the past.

Answer 1

So that you can collect the number of data that you need according to the optimization of costs using the most positive data from the past, I suggest you make a control chart which is done as follows:

They are made using a coordinate system, whose horizontal axis indicates the time the data is framed, while the vertical axis serves as a scale to transcribe the measurement made. Measurement points are joined by straight lines. To plot the control chart, set the limits plus / minus the previously set tolerance values ​​for the process being measured.

1. Control chart for fraction of nonconforming units (p) - p is the percentage of nonconforming units found in the controlled sample.

2. Control chart for number of nonconforming units (np) - It is equivalent to the previous chart, but applicable only if all the samples are the same size (n). - np = number of nonconforming units.

3. Control chart of nonconformities per unit (u) - It is used when several independent nonconformities (defects) can appear in the same unit of product or service, examples: assembly of complex components such as televisions, computers or provision of services with multiple points of contact with the client. - u = number of nonconformities of a unit.

4. Chart of control of number of nonconformities (c) - It is equivalent to the previous chart, but applicable only if all the samples are of the same size (n). - This graph is also used when the nonconformities are dispersed in a more or less continuous flow of product. - c = number of nonconformities.