3.4 Limitations and Prospects
Although the performance of the enhanced prediction
model has been relatively improved, there are also
some noticeable limitations that require to be
accommodated in forthcoming researches. One major
limitation is that the model relies on historical stock
price data and technical indicators, which may not
fully capture the impact of unforeseen market shocks
or changes in investor sentiment. ExxonMobil's stock
belongs to the energy sector and is particularly
sensitive to external factors such as crude oil price
fluctuations, geopolitical events, regulatory changes,
and broader economic conditions. These factors can
also lead to sudden and large price fluctuations, which
are difficult to predict by technical indicators alone.
Another notable limitation is the multicollinearity
between the metrics used in the model. The VIF
analysis shows that some indicators, such as the 10-
day SMA and 20-day EMA, have extremely high
values, suggesting that the information provided by
these variables overlaps with each other and can lead
to overfitting. Multicollinearity reduces the predictive
power of the model, making it less robust when
applied to different market conditions. Solving this
problem through feature selection, dimensionality
reduction, or advanced regularization techniques can
help to enhance the stability and universality of the
model.
There are also differences in the model's
performance during periods of high market volatility,
highlighting the challenges of predicting rapid price
changes driven by external events. While the
consolidation of indicators such as the RSI and
MACD adds value by capturing momentum and trend
reversals, these enhancements are still not enough to
fully adapt to sudden changes. Incorporating external
macroeconomic variables, such as real-time oil prices,
global economic indicators, or sentiment analysis
from news sources, can further refine the model and
improve its ability to respond quickly to market
changes.
4 CONCLUSIONS
To sum up, this study aimed to improve time series
forecasting of ExxonMobil’s stock prices by
integrating moving averages with additional technical
indicators using the stockstats Python library. The
enhanced model demonstrated significant
improvements in predictive accuracy, with an R-
square value of 0.8546, highlighting its ability to
closely track actual stock prices. By incorporating
indicators like EMA, RSI, and MACD, the model
provided valuable insights into market trends and
momentum, making it a more effective tool for
financial analysts and investors. However, the study
also highlighted challenges such as multicollinearity
among indicators and the limitations of relying solely
on historical data. Future research should focus on
refining the model by incorporating additional
macroeconomic factors, employing advanced
machine learning techniques, and reducing
multicollinearity to further improve forecasting
accuracy. Overall, this research underscores the
importance of adapting time series models to the
unique characteristics of the energy sector, offering
valuable tools for financial analysis and investment
strategy development.
REFERENCES
Atsalakis, G. S., Valavanis, K. P., 2009. Forecasting stock
market short-term trends using a neuro-fuzzy based
methodology. Expert Systems with Applications, 36(7),
10696-10707.
Bao, Z., Wang, C., 2022. A multi-agent knowledge
integration process for enterprise management
innovation from the perspective of neural networks.
Information Processing & Management, 59(2), 102873.
Bustos, O., Pomares-Quimbaya, A., 2020. Stock market
movement forecast: A systematic review. Expert
Systems with Applications, 156, 113464.
Ding, G., Qin, L., 2020. Study on the prediction of stock
price based on the associated network model of LSTM.
International Journal of Machine Learning and
Cybernetics, 11(6), 1307-1317.
Fischer, T., Krauss, C., 2018. Deep learning with long
short-term memory networks for financial market
predictions. European Journal of Operational Research,
270(2), 654-669.
Guresen, E., Kayakutlu, G., Daim, T. U., 2011. Using
artificial neural network models in stock market index
prediction. Expert Systems with Applications, 38(8),
10389-10397.
Lin, G., Lin, A., Cao, J., 2021. Multidimensional KNN
algorithm based on EEMD and complexity measures in
financial time series forecasting. Expert Systems with
Applications, 168, 114443.
Su, Y., Cui, C., Qu, H., 2022. Self-Attentive Moving
Average for Time Series Prediction. Applied Sciences,
12(7), 3602.
Tsay, R. S., 2019. An introduction to time series analysis
and forecasting with applications to financial data.
Wiley Series in Probability and Statistics.
Yadav, D. K., S., Goswami, L., 2024. Autoregressive
Integrated Moving Average Model for Time Series
Analysis. 2024 International Conference on
Optimization Computing and Wireless Communication
(ICOCWC), Debre Tabor, Ethiopia, 1-6.