marginally higher error rate. Conversely, the
XGBoost model shows the poorest performance in
this classification task, especially with an increased
occurrence of false positives and false negatives,
implying that it may have overfit the model and
resulted in subpar performance on the test data.
The findings indicate that the LR model
demonstrates superior overall performance in the
binary prediction task for stock price movements,
particularly regarding the balance of precision and
recall. In contrast, the SVM model performs better at
identifying majority class samples (negative
samples). However, it performs poorly when dealing
with minority class samples (positive samples) and
may not be suitable for cases of class imbalance. The
XGBoost model, although typically strong in dealing
with complex categorization tasks, performs slightly
less well than LR in this study, which may be due to
excessive model complexity and overfitting. These
results reveal that simple models such as LR may be
more effective for predicting stock price movements,
especially when the dataset is relatively balanced or
when balancing precision and recall is necessary. At
the same time, it also reminds researchers to consider
the attributes of the dataset and the flexibility of the
model when selecting the model to ensure optimal
prediction results.
4 CONCLUSIONS
Stock price forecasts are vital in financial markets,
aiding investors in making informed decisions,
mitigating risks, and enhancing returns on
investment. Understanding and managing market
volatility are also crucial for the stable growth of
financial markets. This research applies three models
- LR, SVM, and XGBoost - to predict Samsung’s
stock price fluctuations. The experimental results
demonstrate that LR provides the best performance in
binary classification tasks, particularly in terms of
balancing accuracy and recall. In contrast, the SVM
model shows proficiency in recognizing the majority
class (negative samples) but struggles with minority
class (positive samples) identification, making it less
effective in situations of class imbalance. The
XGBoost model, typically strong in complex
categorization tasks, slightly underperformed
compared to LR in this study, potentially due to
excessive model complexity leading to overfitting.
Overall, this paper underscores the importance of
model selection in stock price prediction by analyzing
the efficacy of different models. Future studies could
focus on optimizing model selection, exploring more
sophisticated and diverse data to improve the
accuracy and reliability of forecasting. The findings
of this study offer empirical evidence supporting
intelligent forecasting in financial markets and
suggest new directions for advancing machine
learning models within the financial sector.
REFERENCES
Cao, J., Lin, Z., 2015. Extreme learning machines on high
dimensional and large data applications: A survey.
Mathematical Problems in Engineering, 2015(1),
103796.
Carmona, P., Dwekat, A., Mardawi, Z., 2022. No more
black boxes! Explaining the predictions of a machine
learning XGBoost classifier algorithm in business
failure. Research in International Business and Finance,
61, 101649.
Jiang, Y. Q., Cao, S. E., Cao, S., Chen, J. N., Wang, G. Y.,
Shi, W. Q., Chen, G. H., 2021. Preoperative
identification of microvascular invasion in
hepatocellular carcinoma by XGBoost and deep
learning. Journal of Cancer Research and Clinical
Oncology, 147, 821-833.
Kumar, G., Jain, S., Singh, U. P., 2021. Stock market
forecasting using computational intelligence: A survey.
Archives of Computational Methods in Engineering,
28(3), 1069-1101.
Liu, C., Wang, J., Xiao, D., Liang, Q., 2016. Forecasting
S&P 500 stock index using statistical learning models.
Open Journal of Statistics, 6(6), 1067-1075.
Usmani, S., Shamsi, J. A., 2021. News sensitive stock
market prediction: Literature review and suggestions.
PeerJ Computer Science, 7, e490.
Wang, J., Cheng, Q., Dong, Y., 2023. An XGBoost-based
multivariate deep learning framework for stock index
futures price forecasting. Kybernetes, 52(10), 4158-
4177.
Yang, C., Chen, M., Yuan, Q., 2021. The application of
XGBoost and SHAP to examining the factors in freight
truck-related crashes: An exploratory analysis.
Accident Analysis & Prevention, 158, 106153.
Yelne, A., Theng, D. 2021. Stock prediction and analysis
using supervised machine learning algorithms. In 2021
International Conference on Computational
Intelligence.
Zhou, S., 2021. Sparse SVM for sufficient data reduction.
IEEE Transactions on Pattern Analysis and Machine
Intelligence, 44(9), 5560-5571.