Machine Learning Algorithms General steps

1, using machine learning to solve the problem, we use mathematical language to describe it, and then build a model, such as regression model or classification model to describe the problem;

2, by minimizing the error, maximum likelihood, maximum posterior probability and so on to establish the cost function of the model, to transform to the optimization problem. Find the solution to the optimization problem, which is the best model parameter to fit our data;

3, solve the cost function, find the optimal solution.

To find the best solution method:

1, if the optimization function exists analytic solution. For example, we find that the best value is usually the derivation of the optimization function, the point of the derivative is 0. If the cost function can be simply derivative, and there is an analytic solution to the equation of 0 after derivation, then we can get the optimal parameter directly.

2, if the formula is very difficult to derivative, such as the existence of implicit variables or variables in the existence of coupling, mutual dependence of the situation. Or after the derivation of the formula can not be interpreted, or the number of unknown parameters is greater than the number of equations. At this point, the iterative algorithm is used to find the optimal solution step-after-step.

In particular, if the optimization function is a convex function, then there is a global optimal solution, if the function is non-convex, then there will be many local optimal solutions, so the importance of convex optimization is self-evident. People always want to find the best of him in a myriad of things.

Machine Learning Algorithms General steps