藉由交替K均值分群程序對潛在群體做預測

Abstract

潛在群體模型中的參數估計可以利用群體分析的方法。但是在高維度的資料下，群體分析中的變數選擇問題就更顯重要。這裡，我們發展了群體方法中的交替k均值分群，並利用此想法先找出干擾變數以及此模型中最佳的潛在族群體分群，再利用分群後的結果對其他參數做估計。我們並針對潛在群體模型創造一種分類規則。這個程序不但能完整的分析複雜疾病，對於基因資料分析中也有極大的幫助。我們將分析實際的資料來證實這種方法的優點。

Parameters in latent class analysis could be estimated by some clustering methods. But in the high-dimensional data, variable selection in cluster analysis is an important problem. Here, we propose an alternate k-means clustering method to first distinguish clustering and noisy surrogates and then estimate the parameters in the latent class model. We also create a classification rule, based on the finite mixture model. This classification procedure can explicitly recognize the heterogeneous nature of the complex disease, which makes it perfect in analyzing high-throughput genomic data. The real data analysis demonstrates the advantages of our proposed methods.