An Extended Multivariate Data Visualization Approach for Interactive Feature Extraction from Manufacturing Data

摘要

Awareness about interconnectivities and interactions among parameters is vital for the identification of the optimal manufacturing routes and economic factors within a manufacturing system. Within this context, multidimensional data projection methods, Principal Component Mapping (PCM) and Sammon's Mapping, have been scrutinized for visualizing multivariate interaction patterns. As a new approach, these techniques were employed in such a way that interactive multi-layer maps could be created. Each layer within the generated map matches to a specific attribute and characteristic of the dataset. Individual layers within the map can be interactively selected and superimposed to show multiple and partial interactions.