A novel adaptive methodology for removing spurious components in a modified incremental Gaussian mixture model

Regarding the computational complexity of the update procedure in the fast incremental Gaussian mixture model (FIGMM) and no efficiency for removing the spurious component in the incremental Gaussian mixture model (IGMM), this study proposes a novel algorithm called the modified incremental Gaussian mixture model (MIGMM) which is an improvement of FIGMM, and a novel adaptive methodology for removing spurious components in the MIGMM. The major contributions in this study are twofold. Firstly, a more simple and efficient prediction matrix update, which is the core of the update procedure in the MIGMM algorithm, is proposed compared to that described in FIGMM. Secondly, an effective exponential model (p(Thv)) related to the number of output components generated in MIGMM, combined with the Mahalanobis distancebased logical matrix (LM), is proposed to remove spurious components and determine the correct components. Based on the highlighted contributions, regarding the removal of spurious components, comparative experiments studied on synthetic and real data sets show that the proposed framework performs robustly compared with other famous information criteria used to determine the number of components. The performance evaluation of IGMM compared with other efficient unsupervised algorithms is verified by conducting on both synthetic and real-world data sets.

Automated summary

In this study, a modified incremental Gaussian mixture model (MIGMM) algorithm is proposed as an improvement of FIGMM, along with an adaptive methodology for removing spurious components in MIGMM. The contributions include a more simple and efficient prediction matrix update compared to FIGMM, and the use of an effective exponential model and logical matrix to remove spurious components. Experimental results demonstrate the robust performance of the proposed framework in comparison to other methods.