Teaching Reform of Digital Signal Processing Driven by Probabilistic Neural Network

Abstract

In this paper, teaching reform of Digital Signal Processing (DSP) is conducted based on the data analysis by machine learning methodologies. First, the Spearman correlation coefficients between different process assessments and the total scores are computed to show their relevance. Then, a probabilistic neural network is trained based on real data, and the test result proves that one student’s final score level can be roughly inferred based on his/her process assessment. Hence, several reform schemes are proposed centering at the process assessment to help strengthen learning of important knowledge points. Finally, the assessment results show that the proposed teaching reform plans are reasonable and effective in improving the achievability of DSP teaching.