Lists
Lists
Almada, Arlindo (2021) A proactive chatbot framework designed to assist students based on the PS2CLH model. Doctoral thesis, London Metropolitan University.
Nowadays, universities are using new technologies to improve the efficiency and effectiveness of learning and to assist students to enhance their academic performance. In fact, for decades, new ways to convey the information required to teach and support students have slowly been integrated into education. This development started decades ago with the popularity of e-mails and the Web.
A review of relevant literature revealed that learning requires more innovative and efficient technologies to cope with natural learning challenges, highlighting a need for more effective tools to establish the interaction between humans and machines, lecturers and students. In addition, the covid pandemic presented additional new challenges for the collaboration/interaction of lecturers and students at universities. This situation led to a great demand for such tools. Researchers have been trying to develop such tools for decades, and have made good progress, but they are still in their infancy. There has been a significant evolution in computer hardware in the last decade, leading to advances in AI machine learning and Deep Learning which have made tools such as chatbots more usable. However, the efficiency and effectiveness of the chatbot are still insufficient to meet many educational needs. According to our investigation, current chatbots are mainly based on subject knowledge and therefore assist users with answers which take no consideration of their personal circumstances, which is essential in education.
This research aims to design a proactive chatbot framework to assist students. The new chatbot framework integrates students’ learning profiles and subject knowledge, making the chatbot more intelligent to improve student learning and interaction more effectively. The research consists of two main parts. The first part seeks to determine the most effective students’ learning profiles on the basis of the controllable academic factors which affect their performance. The second part develops a chatbot framework to which students’ learning profiles will be applied. Due to the different nature of these two endeavours, a hybrid methodology was used in this research.
The literature on learners’ characteristics and the academic factors that affect their performance was reviewed in depth, and this formed the basis for developing a new PS2CLH (psychology, self-responsibility, sociology, communication, learning and health & wellbeing) model on which an individual’s web profile can be built. The PS2CLH model combines the perspectives of psychology, self-responsibility, sociology, communication, learning and health & wellbeing to build a student-controllable learning factor model. This study identifies the impact of students’ controllable factors on their achievement. It was found that the model was 94% accurate. In addition, this research raised participant students’ awareness of PS2CLH perspectives, which helped learners and educators to manage the factors affecting academic performance more effectively.
A comprehensive investigation, including a survey, showed that the chatbot supported by AI technology performed better and more efficiently in various assistant situations, including education. However, there is still room for improvement in the effectiveness of the education chatbot. Therefore, the research proposes a new chatbot framework assistant which will integrate students’ learning profiles and develop components to improve student interaction.
The new framework uses knowledge from the PS2CLH model AI - Deep Learning to build a proactive chatbot for assisting students’ learning of their academic subjects and their controllable factors that affects students’ performance. One of the principal novelties of the chatbot framework lies in the communication facilitator between student-lecturer/assistant. The proactive chatbot applies multimodality to the students’ learning process to retain their attention and explain the content in different ways using text, image, video and audio to assist the students and improve their learning experience effectively. Furthermore, the chatbot proactively suggests new controllable factors for students to work on, including related factors that influence their academic performance. Tests of the framework showed that the proactive chatbot demonstrated better question response accuracy than the current BERT (Bidirectional Encoder Representations from Transformers) chatbot and presented a more effective learning method for students.
 |
View Item |