SENIOR: Efficient Query Selection and Preference-Guided Exploration in Preference-based Reinforcement Learning