PENERAPAN ALGORITMA LEARNING VECTOR QUANTIZATION UNTUK PREDIKSI NILAI AKADEMIS MENGGUNAKAN INSTRUMEN AMS (ACADEMIC MOTIVATION SCALE)

Intisari / Abstract

Academic values is an important component for students and colleges. For students, high academic value can facilitate them to find a good job. As for college, academic grades is one of the tools to measure the success of teaching and learning in the college environment. Predicted value of academic achievement by a student is one of the ways that often do colleges to improve the quality of graduates. Learning Vector Quantization method is one of the artificial neural network algorithms that can be used to make predictions. LVQ will fixing weights and output vectors each acquired a new input vector automatically . The amount of training data which are used in this paper are 13 questionnaires and test data which are used 10 questionnaire. The number of variables that will be used as the input vector are 7 factors of motivation.

There are the value of questions that is relating to Self-efficacy, Identification with Academic, Intrinsic motivation, extrinsic motivation, Amotivation, Meaningful Shallow cognitive, cognitive engagement and student engagement as measured by the scale linkert. While the number of output vectors are 4  academic value. That value are less, sufficient, good and satisfactory. The output of this research that using 9 training data, 0,05 learning rate, 100 epoch, 10 test data produce 60% accuracy. This output could change for the better level of accuracy by testing and varying the value of learning rate, epoch and training data. The more data that is used to train the LVQ will have a more complete knowledge.