Are clusterings of multiple data views independent?

Abstract

In the Pioneer 100 (P100) Wellness Project (Price and others, 2017), multiple types of data are collected on a single set of healthy participants at multiple timepoints in order to characterize and optimize wellness. One way to do this is to identify clusters, or subgroups, among the participants, and then to tailor personalized health recommendations to each subgroup. It is tempting to cluster the participants using all of the data types and timepoints, in order to fully exploit the available information. However, clustering the participants based on multiple data views implicitly assumes that a single underlying clustering of the participants is shared across all data views. If this assumption does not hold, then clustering the participants using multiple data views may lead to spurious results. In this paper, we seek to evaluate the assumption that there is some underlying relationship among the clusterings from the different data views, by asking the question: are the clusters within each data view dependent or independent? We develop a new test for answering this question, which we then apply to clinical, proteomic, and metabolomic data, across two distinct timepoints, from the P100 study. We find that while the subgroups of the participants defined with respect to any single data type seem to be dependent across time, the clustering among the participants based on one data type (e.g. proteomic data) appears not to be associated with the clustering based on another data type (e.g. clinical data). Data integration; Hypothesis testing; Model-based clustering; Multiple-view data. *

Two-stage Linked Component Analysis for Joint Decomposition of Multiple Biologically Related Data Sets

“…Park & Lock (2020) jointly analyzed multiple data sets for heterogeneous groups of objects with heterogeneous feature sets. Gao et al (2020) and Wang & Allen (2019) considered clustering problems for multi-view data. Li et al (2018b) proposed a regression model with multi-view data as covariates.…”

“…Another example is in the study of Alzheimer’s disease where recent efforts have been focusing on combining brain imaging data, genetic data, as well as clinical outcomes in predicting disease (Nathoo and others , 2019). A third example is a large study profiling different states of wellness, where genetic, proteomic and metabolic data among other types of data are collected (Gao and others , 2020). Together, the different types of data provide a more comprehensive picture which has the potential of better characterizing and optimizing what it is to be healthy.…”

“…In addition to these examples of studies with multimodal data collection, many public repositories are full of experimental data from single modalities that are related biologically. Therefore, in recent years, there has been growing interests as well as demands for understanding and utilizing these data in an integrative way (Lock and others , 2013; Yang and Michailidis, 2016; Li and Jung, 2017; Gaynanova and Li, 2019; Gao and others , 2020).…”

Intruder Alert: Dimension Reduction and Density-Based Clustering for a Cybersecurity Application

“…As should be clear, the distances may be meaningful, but there is no clear way to say which distances are associated with a group of malicious observations. We might consider testing the difference, but, as Gao, Bien, and Witten highlighted recently [22], this should be considered double-dipping. The clusters are generated expressly to identify a difference.…”

“…In threat detection, we would prefer to capture all malicious traffic, even if that requires more false positives. Gao and colleagues [22] used adaptive control version of PCA for feature extraction alongside an incremental extreme learning machine as part of an IDS. Li et al proposed an anomaly detection framework for network traffic that used PCA to extract features and reduce redundancy.…”