I have a data mining exercise that I find bizarre to solve. I have a data set of 2500 items for employees, having 13 classes for each employee, ranging from age to expertise. I have to use data mining to help a company to choose the right employees from this data set, looking for employees that : "The people should be able to speak English, their salary should not be more than £350,000 and they should have expertise in the investment and financial sector respectively. In addition, they should be willing to relocate to the UK, where ABCBank is based and not have changed more than 5 jobs in the last 5 years, because then ABCBank considers them unreliable. They should be able to close deals above £2 million"
My question is, if I already know all these requirements, isn't it easier to just use a select node and choose all the people that fulfill these and the create a class that says if they are suitable or not, and then use classification.
Or do I have to use classification for each requirement? for e.g using Salary as a target variable to find the ones below 350000, then expertise and so on until I run classification for all the targets and then intersect populations?