Neural-fuzzy with representative sets for prediction of student performance

Abstract

In this paper, a new method for handling the Multi-Input Multi-Output Student Academic Performance Prediction (MIMO SAPP) problem is proposed. The MIMO SAPP aims to predict the future performance of a student after being enrolled into a university. The existing methods have limitations of using a parameter set and an unsuitable training strategy. Thus, the new method called MANFIS-S (Multi Adaptive Neuro-Fuzzy Inference System with Representative Sets) uses multiple parameter sets and a special learning strategy to resolve those weaknesses. Specifically, the idea of multiple parameter sets is to approximate the MANFIS-S model with many meaningful parameters to ensure the performance of system. This is regarded as the representative problem, which is mathematically formulated and theoretically validated. The idea for the special learning strategy is to use global and local training. In global training, a random parameter set is trained from the first to the last record of the database. Each time of training results in a set of parameters. Global training will rectify and achieve a meaningful subset of parameters by the last training process. In local training, there are 2 types of parameters in MANFIS-S namely premise and consequent that are trained by the gradient descent and Particle Swarm Optimization in a hybrid way. Lastly, for a new record in the testing set, Fuzzy K-Nearest Neighbor is used to find which group it belongs to. The proposed MANFIS-S model is experimentally validated against ANFIS, MANFIS, OneR and Random Tree in a benchmark student performance dataset from UCI, a real student performance dataset from VNU University of Science, Vietnam, and 3 educational datasets taken from KDD Cup. The experiments demonstrated the superiority of MANFIS-S over the related algorithms in term of accuracy.

New Soft Set Based Class of Linear Algebraic Codes

“…Further introduction of soft code with rank metric [22] and construction of T-direct soft codes [23] may be helpful to tackle the multichannel coding problem, which is left for researchers in coding theory. The general case based on N-soft sets and others [24][25][26][27][28][29][30][31][32][33][34][35][36] will be developed as well.…”

A new fusion of salp swarm with sine cosine for optimization of non-linear functions

“…Future studies will investigate a new method to accelerate the speed of HSSASCA as well as apply it for solving other constrained nonlinear optimization functions [62][63][64][65][66][67][68][69][70][71][72][73][74][75].…”

On Data Protection Regulations, Big Data and Sledgehammers in Higher Education

“…Higher Education institutions have not been alien to the discussions, as some scholars argued that gathering as much information as possible about university students, professors and administration staff could enable deep analysis and, thereby, proactive actions in student attention, course planning and resource management [4][5][6]. In this line, there have been numerous studies in the recent past about predicting student outcomes using Artificial Intelligence (AI) techniques [7][8][9][10][11][12][13][14][15][16][17] and it is generally assumed that the more abundant the data, the more accurate the predictions.…”

“…Our experiment consisted in performing a scan of hyperparameters for a Multi-Layer Perceptron (MLP) neural network, in search for the configuration that attained the greater accuracy in predicting academic outcomes from the socio-economic data. We chose the MLP for being one of the best understood machine learning models, commonly used in the related literature [18,19]; its best configuration would be used as a benchmark for the comparison of other techniques, including the ones used in References [7][8][9][10][11][12][13][14][15][16][17] and more advanced neural network schemes. However, the scan of hyperparameters revealed no correlations or dependencies between the input variables and the chosen metrics in any case, showing that-at least for the UPS and alike settings-there is no actual gain from applying machine learning techniques on extensive socio-economic data.…”

“…In relation to these precedents, plus those of References [14][15][16][17], our work involves the most extensive and fine-grained dataset of student records (30 times as many as the average from above). Besides, whereas many others have focused on comparing different algorithms and techniques with single configurations or, at the most, a few tens of combinations of hyperparameters, we are the first to conduct an extensive scan (more than 12,000 configurations) to make the most of one technique that ranked well in the literature, namely the Multi-Layer Perceptron.…”