Creativity index calculation with question answering system using BERT model

Abstract

This system proposes a question-answer model to analyze answers given by the users to the open-ended question. For this system, we use a squad dataset. The squad dataset is helpful for open-ended questions. The model uses NLP and the BERT question-answer model for handling the answers. Problem-solving deals with the analysis of questions and answering them in a creative, feasible, and efficient way. The problem-solving the approach gives the way of thinking of the problem solver. Computational methods along with NLP are being developed to calculate these indices. With the help of question-answer systems in NLP, the users’ answers are analyzed and measured are being attempts to perform such analysis but the answers to the questions are varied from person to person which makes answer analysis challenging. Answers given by users and answer for the question present in the squad dataset is compared by using cosine similarity to calculate the creativity index. The creativity index is a linear combination of fluency, flexibility, uniqueness, and elaboration indices; calculated for every given answer by comparing given answers and the stored answers. The system is trained and tested for the general questions present in the squad dataset. This system is beneficial to predict and measure the creative thinking, problem-solving ability, and thinking ability of the user. It helps to build the work culture, societal culture, and behavior of individuals in the institution, community, and workplace. The CI is a tremendously useful tool for businesses, office employees, and students. Thinking and problem-solving skills are calculated using this technique.