Mixture Gaussian process model with Gaussian mixture distribution for big data

• Published in: Chemometrics and intelligent laboratory systems Vol. 253; p. 105201
• Main Authors: Guan, Yaonan, He, Shaoying, Ren, Shuangshuang, Liu, Shuren, Li, Dewei
• Format: Journal Article
• Language: English
• Published: Elsevier B.V 15.10.2024
• Subjects: Big data; Gaussian mixture distribution; Gaussian process; Inducing variables; Machine learning; Big data; Gaussian process; Inducing variables; Gaussian mixture distribution; Machine learning
• ISSN: 0169-7439
• DOI: 10.1016/j.chemolab.2024.105201

In the era of chemical big data, the high complexity and strong interdependencies present in the datasets pose considerable challenges when constructing accurate parametric models. The Gaussian process model, owing to its non-parametric nature, demonstrates better adaptability when confronted with complex and interdependent data. However, the standard Gaussian process has two significant limitations. Firstly, the time complexity of inverting its kernel matrix during the inference process is O(n)3. Secondly, all data share a common kernel function parameter, which mixes different data types and reduces the model accuracy in mixing-category data identification problems. In light of this, this paper proposes a mixture Gaussian process model that addresses these limitations. This model reduces time complexity and distinguishes data based on different data features. It incorporates a Gaussian mixture distribution for the inducing variables to approximate the original data distribution. Stochastic Variational Inference is utilized to reduce the computational time required for parameter inference. The inducing variables have distinct parameters for the kernel function based on the data category, leading to improved analytical accuracy and reduced time complexity of the Gaussian process model. Numerical experiments are conducted to analyze and compare the performance of the proposed model on different-sized datasets and various data category cases. •Non-parametric models are well suited for complex and strongly coupled data.•Gaussian mixture distribution captures essential characteristics as inducing variables of GPM.•Soft classification boundaries are effective for modeling mixed or overlapping features.•The introduction of SVI enhances computational efficiency with parallel computation.•GPM with varying weights of different classes improves modeling performance.