融合随机统计规律与优化思想的成分数据预测方法研究OACSTPCD

Frequency distribution series prediction is an important class of problems in compositional data research.By integrating the statistical regularity of frequency distribution and optimization idea,the prediction method combining grey prediction theory and optimization method are used to construct the mathematical prediction model for quadratic programming by taking the minimization of the difference between the predicted values of expectation and mean as the goal,and setting the combination of historical experience constraints and constituent data constraints as the constraints.The empirical analysis of the frequency distribution structure data of core competencies among students in a certain major are used to verify the the effectiveness of the proposed prediction model.The results show that the average absolute error of the proposed prediction model is 0.0416,the root mean square error is 0.0460,and the directional cosine is 0.9824.In comparison with the spherical coordinate transformation prediction results,the proposed model has better prediction accuracy and is feasible.The research results show that the proposed prediction model has better prediction accuracy and can effectively solve the problem of frequency distribution sequence prediction,and meanwhile it can provide a new research idea for the prediction problem of compositional data containing 0 or 1 in the compositional data.