Balancing Exploration and Exploitation in Chatbot Training Using Reinforcement Learning Algorithms