has higher complexity. At the same time, the higher
complexity of the model will strengthen the burden
on calculation and the computer. XGBoost needs a
great amount of storage and resources for calculation,
this may cause overheat of the computer due to a long
period and high-intense workload if the hardware is
not perfect enough.
Because of the complexity of the model, it is able
to produce a model concluded from the huge amount
of data. However, overfitting might occur if the scale
of the data is not big enough or the noise is too much.
Although normalization and standardization probably
could solve this problem, they need to be debugged
seriously and constantly. It takes a long time and is
challenging but produces similar results to other
machine learning models.
Though with these disadvantages, XGBoost is
still an effective model that could produce quite
complex and accurate results, as it could catch non-
linear relations (Ji et al., 2022; Yaswanth &
Jaisharma, 2024). Hopefully, in the future,
researchers could explore more the adjusting of the
hyperparameters and model optimizing, in order to
increase the precision and accuracy of the model
stability and the model prediction. At the same time,
the author hopes that further majorization of the
model could increase the ability of XGBoost to
perform real-time predictions. Which could lead to
the popularization of the model and better ability to
suit the changes in the market.
4 CONCLUSIONS
This passage built up a predictive model applied in
gold price prediction based on the XGBoost
algorithm. XGBoost makes use of decision trees to
achieve regressions. Through the prediction results of
previous data, the intelligent portfolio optimization
model proposed in this paper uses the model to weigh
the transaction cost and the income obtained in the
transaction to decide whether to trade, and combines
the traditional portfolio model with the algorithm of
machine learning to better apply to the portfolio
research. A regression tree is a type of decision tree
used in machine learning that is designed to predict
continuous target variables based on multiple input
features. It uses a tree model that describes decisions
and their possible consequences, especially target
predictions, to translate observations about a project
into target value conclusions for that project. The
results show that the gold price prediction model
based on the XGBoost algorithm performed
excellently at capturing short-term market
fluctuations and long-term trends. Compared with
traditional time series models, XGBoost can better
deal with complex non-linear relationships and large-
scale feature sets, at the same time, it could improve
the accuracy and stability of the prediction. Although
the XGBoost itself is a black box model, through
feature engineering and data visualization, the
research is able to understand which factors have a
decisive impact on the price of gold. This capability
provides decision-makers with deeper market insight
and helps them develop more precise investment
strategies and risk management programs.
REFERENCES
Amini, A., Kalantari, R., 2024. Gold price prediction by a
CNN-Bi-LSTM model along with automatic parameter
tuning. Plos one, 19(3), e0298426.
Beckmann, J., Berger, T., Czudaj, R., 2019. Gold price
dynamics and the role of uncertainty. Quantitative
Finance, 19(4), 663-681.
Cotty, S. R., Kim, N., Su, X., 2023. Electrochemically
mediated recovery and purification of gold for
sustainable mining and electronic waste recycling. ACS
Sustainable Chemistry & Engineering, 11(9), 3975-
3986.
Erdoğdu, A., 2017. The most significant factors influencing
the price of gold: An empirical analysis of the US
market. Economics, 5(5), 399-406.
Jabeur, S. B., Mefteh-Wali, S., Viviani, J. L., 2024.
Forecasting gold price with the XGBoost algorithm and
SHAP interaction values. Annals of Operations
Research, 334(1), 679-699.
Ji, S., Wang, X., Lyu, T., Liu, X., Wang, Y., Heinen, E.,
Sun, Z., 2022. Understanding cycling distance
according to the prediction of the XGBoost and the
interpretation of SHAP: A non-linear and interaction
effect analysis. Journal of Transport Geography, 103,
103414.
Lili, L., Chengmei, D., 2013. Research of the influence of
macro-economic factors on the price of gold. Procedia
Computer Science, 17, 737-743.
Qian, Y., Ralescu, D. A., Zhang, B., 2019. The analysis of
factors affecting global gold price. Resources Policy,
64, 101478.
Shi, J., 2023. Examine Bitcoin price predictability with
machine learning algorithms. In Third International
Conference on Intelligent Computing and Human-
Computer Interaction (ICHCI 2022) (Vol. 12509, pp.
621-628).
Yaswanth, V., Jaisharma, K., 2024. Prediction of thyroid
nodules using novel optimized XGBoost machine
learning algorithm based on its size compared with non-
linear machine learning algorithm. In AIP Conference
Proceedings (Vol. 2853, No. 1). AIP Publishing.