Academic Journal of Mathematical Sciences, 2024, 5(3); doi: 10.25236/AJMS.2024.050307.

## Gaussian Process Regression Model Based on Multi-Kernel Learning and Bagging Algorithm

Author(s)

Jiaru Wang1, Zhenhan Chen2

Corresponding Author:
Jiaru Wang
Affiliation(s)

1School of Applied Mathematics, Shanxi University of Finance and Economics, Taiyuan, 030000, China

2School of Economics, Xinjiang University of Finance and Economics, Urumqi, 830000, China

### Abstract

The Gaussian process regression model is a flexible and powerful non -parameter regression method. It can fit the relatively complicated function of the Reproducing Kernel Hilbert Space (RKHS) by selecting the appropriate kernel function. However, the correctness of the selection of the kernel function has an important effect on the efficiency of the predictive model and there is no uniform paradigm for solution. Based on this, this article proposes a new self-adaptive Gaussian process regression model by using multiple kernel function. On the one hand, this proposed method can fit the predicted function in a more large RKHS, and adapts to solve the selection problem of kernel function. On the other hand, this method enhances the stability of the Gaussian process regression model by using the Bagging algorithm. In addition, the basis model in the Bagging framework is model-free. The simulation data analysis shows that the proposed model has smaller mean errors, relative errors and absolute errors relative to some comparison methods. Lastly, it is applied to the actual stock data of quantitative transactions and the results also reflect the accuracy and stability of the proposed model.

### Keywords

Stock Price Modeling, Gaussian Process Regression, Multi-Kernel, Bagging Algorithm

### Cite This Paper

Jiaru Wang, Zhenhan Chen. Gaussian Process Regression Model Based on Multi-Kernel Learning and Bagging Algorithm. Academic Journal of Mathematical Sciences (2024) Vol. 5, Issue 3: 44-51. https://doi.org/10.25236/AJMS.2024.050307.

### References

[1] Wang S, Mo D, Wu Q, et al. Design and analysis of sustainable models for Qinling ecological protection and mining development[J]. Minerals Engineering, 2023, 204: 108446.

[2] He Y , Liu J N K , Hu Y ,et al.Fuzzy Integral-Based Kernel Regression Ensemble and Its Application[J].handbook of research on artificial intelligence techniques & algorithms, 2015.DOI:10.4018/978-1-4666-7258-1.ch012.

[3] Jones, O., Scheuerlein, A., Salguero-Gómez, R. Et Al. Diversity of Ageing across the Tree of Life. Nature 505, 169–173 (2014).

[4] Shu Shike, Li Lu. Multi -factor stock selection strategy based on Elastic Net punishment [J]. Statistics and decision -making, 2021 (16): 5.

[5] Yan Zhengxu, Qin Chao, Song Gang. Random forest model stock price prediction based on Pearson feature selection [J]. Computer engineering and application, 2021, 57 (15).

[6] Tian, L., FENG, L., yang, L. Et Al. Stock price prediction based on LSTM and LightGBM hybrid models. Super Computing Magazine 78, 11768–11793 (2022).

[7] Breiman Leo. Random Forests [J]. Machine Learning, 2001, 45 (1): 5-32.

[8] Wang Labin, Li Chao, Yang Qinmin, et al. A Wind Speed Vector Prediction Method Based on Gaussian Process and Combined Model: CN202211356272.6[P]. CN115907099A [2024-03-25].

[9] Yang Zhenli, Xia Kewen. The stock price prediction model regressioned based on the Gauss process [J]. Computer simulation, 2013, 30(1):5.

[10] Zhu Hongyu Gauss's process model prediction of stock prices [D]. Yunnan University, 2017.

[11] He Zhongzheng, Fang Li, Liu Wan, etc. The short-term runoff prediction research based on the index kernel function Gauss process [J]. China's rural water conservancy and hydropower, 2023 (8): 25-31.

[12] Zhu Shuxian, Zhang Renjie. Research on the selection of the kernel function of the vector machine [J]. Science and Technology and Engineering, 2008, 8(16): 5.

[13] Sun Na, Zhang Nan, Zhang Shuai, et al. A monthly runoff forecast model based on the Gaussian process of the Gaussian process and its application [J]. Hydropower and Energy Science, 2023.

[14] Yang Mengmeng, Wang Xinghui, Zhao Xing. Stock price forecast based on the average method of enhanced the model based on dynamic factor [J].Journal of Qilu University of Technology, 2022, 36(5):75-82.

[15] Qiu Shan, Gong Wenjie, Zhang Zhisheng. Based on Bagging-short-term load prediction model study based on the Bagging-combination kernel function related vector machine [J].Journal of Electrical Engineering, 2023, 18(2):142-148.

[16] Niu Hongli, Zhao Yazhi. Use the Bagging algorithm and GRU model to predict the stock price index [J]. Computer engineering and application, 2022, 58 (12): 132-138.