Publication

Abstract : Chatbots are existing since few years and recently it has started acquiring popularity. Earlier to chatbots, people use help desk as the enquiring medium and hence people working at help desks have to work all the days and answer all the questions. Most of the queries are repetitive in nature and answers are given from a structured database. In order to reduce the effort of humans, we can have a chatbot deployed for the same activity. This work focuses on a chatbot which has been developed to provide a faster human-like interaction for admission enquiry system. The chatbot is capable of handling negative or irrelevant scenarios and responds to the queries in faster manner. Decision making by the chatbot on choosing the right set of sentences is done using LSA algorithm and cosine similarity. In addition to answering, the chatbot also maintains data of questions which is not being answered. This data can be used for future analysis for retrieval-based system. The chatbot also takes the feedback from the customers and this data can be analyzed using the feedback category report generated by the chatbot using LDA algorithm.