Extended Ant Colony Algorithm Based Learning Path Recommendation In The Online Customer Service

Online customer service together with traditional customer service methods had been more and more applied by enterprises. But the key factor that influences the effect of it is how to generate customers’ questions oriented solutions more effectively in online customer service. The research of learning path in online learning provides a research basis. But customers own different character with learners, and they have different purpose compared with learners in online learning, customers in online customer service platform mainly focus on solving problems while learners in online learning intend to complete learning tasks. On the basis of research about learning path in online learning field, the paper tried to find out a more effective way to organize knowledge for customers’ problems, and recommend personalized learning path for customers for the purpose of solving customers’ questions. The core research problem is how to find an optimal learning path from complicated learning objects network that can solve customers’ questions and fit customers’ character as well.The thesis reviewed learning path recommendation related research, analyzed customers’ character and constructed character model. Explored the method to obtain learning objects, and in the further discussed ant colony algorithm based learning path recommendation method, proposed online customer service system framework on the basis of learning path recommendation. Learning objects were acquired by simplified TF*IDF method, and analytic hierarchy process was employed to obtain "ability value" which stand for how much degree the learning objects solve the problem. Ant colony algorithm was used to generate learning path. When in the process of recommendation decision, ant colony algorithm considered the customers’ attributes, learning objects’ attributes and "ability value", which help to achieve personalized and adaptive recommendation. The simulation experiment proved the possibility and adaptation of the proposed method. The proposed method applies together with other self-service methods, can improve the quality and personalization of online customer service, which can in the further improve online customer service level and customer satisfaction.