i-IF-Learn: Iterative Feature Selection and Unsupervised Learning for High-Dimensional Complex Data
This work addresses the problem of data interpretation and clustering for researchers in bioinformatics and related fields dealing with high-dimensional complex data, representing an incremental advancement in feature selection methods.
The paper tackles the challenge of unsupervised learning in high-dimensional data by proposing i-IF-Learn, an iterative framework that jointly performs feature selection and clustering to recover influential features, resulting in significant performance improvements over classical and deep clustering baselines on gene microarray and single-cell RNA-seq datasets.
Unsupervised learning of high-dimensional data is challenging due to irrelevant or noisy features obscuring underlying structures. It's common that only a few features, called the influential features, meaningfully define the clusters. Recovering these influential features is helpful in data interpretation and clustering. We propose i-IF-Learn, an iterative unsupervised framework that jointly performs feature selection and clustering. Our core innovation is an adaptive feature selection statistic that effectively combines pseudo-label supervision with unsupervised signals, dynamically adjusting based on intermediate label reliability to mitigate error propagation common in iterative frameworks. Leveraging low-dimensional embeddings (PCA or Laplacian eigenmaps) followed by $k$-means, i-IF-Learn simultaneously outputs influential feature subset and clustering labels. Numerical experiments on gene microarray and single-cell RNA-seq datasets show that i-IF-Learn significantly surpasses classical and deep clustering baselines. Furthermore, using our selected influential features as preprocessing substantially enhances downstream deep models such as DeepCluster, UMAP, and VAE, highlighting the importance and effectiveness of targeted feature selection.