Exploring the characteristics of folk music tunes based on data mining model

In this paper, during the design process, some sample data are first extracted, and the distance between data is calculated to obtain the symmetric distance matrix along the diagonal. The model mutual information quantity is introduced, and the homogenization factor is detected after weight calculation. Next, analyze the rule weights based on the normalized mutual information quantity and use information entropy to identify abnormal data. Finally, find the minimum set with frequent item set and mine the fuzzy association rules. To verify the feasibility of the model, the results of exploring the characteristics of ethnic music tunes show that the neutral tone color is thin and the ethnic music melody line is soft in tune when the long and short melody lines are 9% and 12%, respectively. The number of strong and weak beats in the beat is 2-7 times, mixed with 2/4 and 3/8 beats alternately, and the musical tune is regular. The third-order interaction of wave amplitude variance is significant, and the two media of CD and phonograph help listeners understand and remember the music. Based on this, the digital model can improve the reproduction effect of folk music and can objectively target the transmission of folk music with national cultural characteristics.