Smart Chatbot System for Banking using Natural Language Processing Tools
Résumé
Abstract This research work focuses on creating a Smart Chatbot System for banking that leverages Natural Language Processing Toolkit and Machine Learning. The study involves gathering and preprocessing a dataset of user queries and responses. Key words are defined for pattern matching, and a decision tree algorithm is used to train the model. The Chatbot's performance is evaluated, and the results demonstrate the successful development of a functional Chatbot with a high F1 score of 0.97, indicating 97% accuracy in understanding user queries and providing relevant responses. User testing and a comparison with an existing system reveal significant improvements achieved through the incorporation of NLP and machine learning algorithms. The Chatbot is designed with an intuitive and user-friendly interface accessible through a web application, and administrators have a separate back-end access to effectively manage the knowledge base. The research also identifies areas for future improvement, such as enhancing accuracy, incorporating system memory and anaphora resolution, and integrating a logging system. Addressing these limitations will contribute to the advancement and effectiveness of conversation agents in the banking sector.
Citer ce document
Accès au document
Texte intégral en lecture en ligne, réservé aux abonnés SPHAERO et aux membres de l'institution. Se connecter
Voir l'article sur le site de la revueAuteur(s)
Statistiques
Consultations : 1
Téléchargements : 0