As developments in educational technology continue to advance, the methods in which the courses are delivered and received by learners' evolved from board teaching to online courses. It is necessary to investigate and understand the progression and advancements in the quality of online education through the learning behaviour. The traditional way of using pre-post test scores and learners' feedback is not much helpful in assessing the online learning behavior. In the proposed research work, a learner's behavior prediction model is to be built using hybrid deep learning techniques. For this model building, various information such as facial expressions, time series Electroencephalography (EEG) data, clinical parameters like pulse rate, blood pressure, and skin temperature are acquired directly from the learner through sensors and other general information namely age, course, gender, location are included. Using this data collection, a hybrid model using Convolutional Neural Network (CNN) and Recurrent Neural Networks (RNN) is to be built and validated in real time for predicting the cognitive ability of a learner.