There are basically three ways of approaching epistemology.
A belief type epistemology is an attempt to describe why people think that some of their judgments are knowledge. The traditional definition of knowledge (true, justified, belief) could be seen as part of this type of investigation, i.e. it isn’t a description of what knowledge is, but what makes people think that they have it.
A truth type epistemology is an attempt to describe which procedures or methods of thinking can best discover truths about the world, or at least identify which statements are more likely to be true. For example a philosopher conducting this type of investigation might endorse Bayes’ theorem, because it seems to accurately reflect how evidence affects the probability that a statement is true, and not because people actually use that theorem (unconsciously), which is what a belief type epistemology would require.
A practical epistemology is an attempt to reveal which truth-finding procedures or methods will be reasonable to use in everyday life. A philosopher conducting this type of investigation would probably reject Bayes’ theorem as a good method, because it is too “costly” to use in some cases (or at least more than the answer is worth). However neither is the philosopher looking for methods that support our intuitions, simply methods that don’t require more calculation and investigation than we value having the answer. What is being sought is a method that is usually right and relatively simple.
I personally think that belief type investigations are of little value to philosophy, not because they are worthless, but because they lie more in the domain of psychology. This is probably a sentiment shared by many modern philosophers, who are after the truth and want to surpass our intuitions. I mention it only because it is too easy to fall accidentally into a belief type investigation when you mean to be conducting an investigation of one of the other two types. For example sometimes we are tempted to consider seriously objections in the form “when we apply theory X you endorse to situation Y it results in Z, which quite clearly goes against our sense of what knowledge is.” The appropriate response is of course to argue that they have simply found a case in which our intuitions are faulty, unless Z really is a bad conclusion (by objective standards).
The choice between the practical and truth type epistemology seems largely a matter of personal preference, as both have their uses. I usually approach epistemology as a truth type investigation, because I am interested in finding methods that lead us to the truth with the most accuracy, irrelevant of the cost, in order to better solve other problems in science and philosophy, where the “cost” of finding a solution is largely irrelevant. Also having a system of truth type epistemology is a great way to rebut solipsists and relativists.