When we have a machine learning model to be built for a use case ,one of the steps is to choose an algorithm. An algorithm has to be chosen out of the various available machine learning algorithms like Decision Tree, Random Forest,Naive Bayes,Timeseries,Neural Networks etc. How is ML algorithm chosen when most can be used for both classification as well as regression use cases ?
For instance ,for our use case we see whether it is a classification or a regression problem. If it is regression then whether it is linear or logistic regression .
While this may be easy to determine, next we need to choose an algorithm for it .Suppose out use case is a case of classification problem. Then decision tree may be a choose for classification problems but then decision trees are used for regression as well. Or say our case is of regression and we may choose a Random Forest but this algorithm is used in case of classification use cases as well.
How is the selection done and how does determining whether the use case is of Regression or Classification help ?