To populate Scholarly, sign in here .


Title Prediction and Classification of Student’s Final Grade Results Based on Reactions and Performance Data
Posted by Orven Llantos
Authors Llantos, Orven; Bandola, Dareen Mae; Fuerzas, Cristy Loraine; Ramos, Jenny
Publication date 2017/10
Conference 14th DOST-ERDT Conference/2017 AEESEAP Workshop & 27th Excutive Committee Meeting/ICSEE International Conference on Sustainable Energy Ecosystems
Abstract The new evolution in the internet world has allowed users to express reactions towards Social Networking Sites (SNS). Students became more vocal about their frustrations and other reactions on SNS compared to face-to-face interactions. Allowing students to also express their reactions towards their performance items and capturing these reactions might be an essential set of data for further student learning. The proponents generated controlled data to represent assumed reactions of students towards their grades. A set of noise data is also generated to accommodate other reactions possibly expressed towards a certain grade. The set of data is then used to associate and classify student’s final grade results based on their reactions and previous grades, in which it was known that affective elements (i.e feelings, reactions) are essential to observe the cognitive ability of the students. Frequent Pattern Growth (FP-Growth) and the Support Vector Machines (SVM) algorithms are used to discover association rules in the periodic grades of the students and to classify their final grading status. The FP-Growth’s performance has a minimum accuracy score of 0.60 - 0.90, as measured by the confusion matrix. The SVM classification on the other hand gave an average score of 0.75 - 0.90 in accuracy. With the high accuracy scores given by FP-Growth and SVM, it was concluded that reactions of the students can be used as additional factors in identifying their final grade results.