An Iterative Multilayer Unsupervised Learning Approach for Sensory Data Reliability Evaluation

This paper investigates the problem of extracting actionable patterns/models from unlabeled and potentially erroneous datasets in an unsupervised way. To address the need for both model extraction and data reliability evaluation, we propose a novel iterative multilayer micro-macro (IM3) method that defines data reliability, learns micro-macro models, and iteratively refines learned models. The IM3 method includes a general data reliability definition to evaluate the reliability level of each sample, a micro-macro model complexity determination, and an iterative data reliability and model complexity update mechanism to overcome the underfitting and overfitting issue. In particular, we propose a consistency-index-based approach to address underfitting and overfitting in an unsupervised way. The refinement of the learned models is enabled via dropping the most unreliable data until the data reliability is above a given threshold. The sensitivity of the proposed IM3 method with respect to the reliability threshold selection is further quantified via false alarm and missdetection to facilitate the selection of an appropriate reliability threshold. Evaluation of the proposed method and quantitative analysis of its sensitivity are provided on a polynomial regression problem via Monte Carlo simulations.