bicycles as a viable transportation alternative may
help increase bicycle sales.
Age was identified as one of the most important
predictors in the RF and DT models. Older
individuals were more likely to purchase bicycles, a
finding that may be explained by health
considerations. As people age, they often become
more concerned with maintaining physical health,
and cycling is increasingly viewed as an accessible
and beneficial form of exercise. From a practical
perspective, marketing campaigns targeting older
demographics that emphasize the health benefits of
cycling could further stimulate bicycle sales in this
segment.
4.3 Key Factors in Bicycle Purchase
Behavior
The results of this study highlight the key factors
influencing bicycle purchase behavior as identified
by the four models. Age, income, and car ownership
consistently emerged as significant predictors
regardless of the model used. These insights provide
valuable information for businesses seeking to target
potential customers, as well as policymakers looking
to promote cycling as a sustainable transportation
option.
For example, RF and XGBoost both identified age
as a crucial factor, indicating that older individuals
are more likely to purchase bicycles. This finding
suggests that marketing strategies could be designed
specifically to appeal to older demographics by
emphasizing the health benefits of cycling and
positioning bicycles as an effective means of
maintaining physical well-being.
4.4 Policy and Business Implications
Similarly, income played a vital role in predicting
bicycle purchases across all models. Higher-income
individuals are more likely to buy bicycles, possibly
due to their greater financial capacity to invest in
high-quality bicycles and a lifestyle that promotes
sustainable transportation. Businesses could focus on
premium pricing strategies or offer advanced bicycle
features targeting affluent consumers.
Car ownership, which showed an inverse
relationship with bicycle purchases, highlights an
area of potential policy intervention. Reducing car
dependency and offering incentives for cycling - such
as building more cycling infrastructure or providing
tax breaks for bicycle purchases - could increase
bicycle sales. This insight helps policymakers create
effective strategies for promoting healthier, more
sustainable urban environments.
5 CONCLUSIONS
The rapid rise in environmental awareness and the
promotion of healthier lifestyles have significantly
contributed to the increase in bicycle purchases
globally. Understanding the factors driving consumer
decisions to purchase bicycles is crucial for both
theoretical research and practical applications, as it
offers valuable insights for businesses and
policymakers alike. This study applied four machine
learning models—RF, DT, XGBoost, and LR—to
analyze bicycle purchase behavior. Each model
demonstrated unique advantages, with RF and
XGBoost showing the most robust predictive
performance. Key features such as age, income, and
car ownership were identified as significant
predictors, reflecting real-world consumer trends and
socioeconomic factors. This study is not without
limits, though. Initially, the models employed were
restricted to a particular group of variables, which
might not adequately represent the intricacy of
customer behavior. Additionally, the data used for
model training may not account for regional or
cultural differences in bicycle purchasing patterns.
These limitations suggest that further research should
incorporate additional variables, such as lifestyle
choices, environmental policies, and urban
infrastructure, to better understand consumer
behavior. Future studies should investigate hybrid
models, which combine the best features of various
machine learning approaches to increase
interpretability and accuracy. Expanding the scope of
analysis to include broader socioeconomic and
cultural factors will allow for a more comprehensive
understanding of bicycle purchase behavior.
Additionally, investigating the role of infrastructure
development, such as bike lanes and public cycling
initiatives, could provide further insights into how
urban planning can promote cycling as a primary
mode of transportation.
REFERENCES
Agrawal, S., Maurya, A. K., 2020. A comparative analysis
of XGBoost. International Journal of Scientific &
Technology Research, 9(2), 132-136.
Chen, T., Guestrin, C., 2020. XGBoost: A Scalable Tree
Boosting System. ACM Transactions on Knowledge
Discovery from Data, 14(1), 1-24.