Page 152 - FULL REPORT 30012024

6.1.1 First Objective

The comprehensive analysis of existing stroke risk assessment methods
identified significant shortcomings, such as a lack of personalization, time-

consuming processes, and limited accessibility. This study underscored the

promise of machine learning algorithms in overcoming these challenges. By
incorporating a wider range of variables, including lifestyle factors and

genetic predispositions, these algorithms can provide a more complete and
personalized risk assessment.

Throughout the research, several obstacles were encountered, particularly in

selecting the most appropriate journals for sourcing information and
determining the best machine learning models and software for the study.

These challenges were successfully navigated during the planning phase,

where an extensive literature review was conducted. This preparatory stage
was crucial in laying the groundwork for the research, enabling the

identification and application of the most suitable machine learning models

and tools to address the identified gaps in stroke risk assessment.

6.1.2 Second Objective

This objective focused around creating an innovative dashboard system and
a tool powered by Random Forest algorithms. This system's design

prioritized both substantial contributions to target users and user-friendliness.
Its intuitive interface was meticulously crafted to demystify complex data,

rendering it accessible and actionable for users. This approach marked a
significant advancement in broadening the accessibility of sophisticated

healthcare technology.

The development journey, however, encountered its share of obstacles.

During the model training phase, several errors arose, particularly when
processing certain user inputs for prediction, which resulted in errors. To

overcome these challenges, the model underwent retraining with an enhanced

focus on feature selections. This step involved refining and selecting relevant
135