Comparing machine learning clustering with latent class analysis on cancer symptoms' data

摘要

Symptom Cluster Research is a major topic in Cancer Symptom Science. In spite of the several statistical and clinical approaches in this domain, there is not a consensus on which method performs better. Identifying a generally accepted analytical method is important in order to be able to utilize and process all the available data. In this paper we report a secondary analysis on cancer symptom data, comparing the performance of five Machine Learning (ML) clustering algorithms in doing so. Based on how well they separate specific subsets of symptom measurements we select the best of them and proceed to compare its performance with the Latent Class Analysis (LCA) method. This analysis is a part of an ongoing study for identifying suitable Machine Learning algorithms to analyse and predict cancer symptoms in cancer treatment.