Bootstrap–CURE: A Novel Clustering Approach for Sensor Data—An Application to 3D Printing Industry

Abstract

The agenda of Industry 4.0 highlights smart manufacturing by making machines smart enough to make data-driven decisions. Large-scale 3D printers, being one of the important pillars in Industry 4.0, are equipped with smart sensors to continuously monitor print processes and make automated decisions. One of the biggest challenges in decision autonomy is to consume data quickly along the process and extract knowledge from the printer, suitable for improving the printing process. This paper presents the innovative unsupervised learning approach, bootstrap–CURE, to decode the sensor patterns and operation modes of 3D printers by analyzing multivariate sensor data. An automatic technique to detect the suitable number of clusters using the dendrogram is developed. The proposed methodology is scalable and significantly reduces computational cost as compared to classical CURE. A distinct combination of the 3D printer’s sensors is found, and its impact on the printing process is also discussed. A real application is presented to illustrate the performance and usefulness of the proposal. In addition, a new state of the art for sensor data analysis is presented.

Bootstrap-CURE Clustering: An Investigation of Impact of Shrinking on Clustering Performance

“…In the earlier research, the authors [13,6] proposed a modification of CURE clustering using several bootstrap samples, however, the shrinking step from the original CURE algorithm is skipped.…”

“…Let us consider a multivariate numerical dataset, with the information about a set I of N, k-dimensional objects as i 1 , i 2 ,...i N . The Bootstrap-CURE proposal [13] begins with drawing a small representative sample of ratio r from the original dataset, and divides it into S bootstrap samples of same size n s without replacement. Each bootstrap sample is subject to hierarchical clustering individually and local clusters are obtained.…”

“…The dataset used in the earlier research [13] has been continued for the current experiments. The data represent a collection of time-series-based sensor data from eight anonymous 3D printers with over 300 printing jobs.…”

“…The final working data after preprocessing contains 46821 records for 41 features. A list of features can be seen in the previous paper [13].…”

“…Guha et al proposed CURE (Clustering-Using-Representation) that scales hierarchical clustering to a large scale by incorporating a sampling strategy. In [13], Suman et al proposed a new algorithm, Bootstrap-CURE that scales hierarchical clustering further by using several bootstrap samples in a CURE like strategy except the shrinking step. In this paper, the impact of shrinking and other related hyper-parameters are investigated on the overall clustering procedure in the context of the original dataset from 3D printing.…”