基于NSGA2的参数自确定数据补全模型

Data Completion Model with Self-Determined Parameters Based on Evolutionary Algorithm NSGA2

基于分解的数据补全模型在补全缺失元素问题的研究中被广泛应用。然而,参数低秩与参数最大迭代次数作为模型的输入,其合理性直接影响补全模型的性能。参数低秩设定不合理将导致数据补全模型出现过拟合或者欠拟合问题。此外,参数最大迭代次数选取不合理将导致计算资源的浪费或者数据补全精度的下降。基于此论文提出一种基于进化算法NSGA2的参数自确定数据补全模型。该模型通过构建多目标函数执行遗传进化操作确定合理的参数值,确保数据补全模型的性能。对比试验结果表明,该模型通过进化算法确定合理参数值有效避免了过拟合与欠拟合问题的发生,同时也避免了计算资源的浪费,确保了数据补全结果的精度。

The data completion model based on decomposition is widely used in the research on the problem of complementing missing elements.However,the rationalities of the low rank parameter and the maximum iteration number parameter as the inputs of the model directly affect the performance of the completion model.Setting an unreasonable low rank parameter will lead to over-fit-ting or under-fitting problems in the data completion model.In addition,selecting an unreasonable maximum iteration number pa-rameter will result in a waste of computing resources or a decrease in the accuracy of data completion.Therefore,this paper propos-es a new data completion model with self-determined parameters based on evolutionary algorithm NSGA2.This model ensures the performance of the data completion model by constructing a multi-objective function and performing genetic evolution operations to determine reasonable parameter values.Comparative experimental results show that the model can determine reasonable parameter values through the evolutionary algorithm,which effectively avoids the occurrence of over-fitting and under-fitting problems,while also avoiding the waste of computing resources and ensuring the accuracy of data completion results.