Page 140 - FULL REPORT 30012024

Table 5.2 Result of Evaluation Metrics

Metrics Result
Accuracy 80%

Precision 75%

Recall 90%

F1 Score 82%

The evaluation metrics of the prediction model yielded notable results,
providing insight into its predictive capabilities. The model achieved an

accuracy rate of 80%, indicating that it correctly predicted stroke outcomes

80% of the time. In terms of precision, the model scored 75%, meaning that
among all the positive predictions for stroke, 75% were accurate. The recall

metric was particularly high at 90%, signifying that the model was

successful in identifying 90% of all actual positive stroke cases. Lastly, the
F1 score, which serves as a balance between precision and recall, was

calculated to be 82%. This score reflects the model's overall efficiency in
minimizing both false positives and false negatives, underscoring its

effectiveness in stroke risk prediction.

5.2 User Acceptance Testing (UAT)

The web-based application for stroke risk assessment was evaluated using

User Acceptance Testing (UAT) and the Technology Acceptance Model
(TAM) as essential approaches. UAT was used to directly evaluate the

feasibility and user-friendliness of the application in real-life situations,
guaranteeing that the final product meets the anticipated requirements and

preferences of its target users. The practical testing method facilitated the
collection of specific comments on the application's performance and user

interface. Conversely, the TAM framework offered a theoretical foundation

for examining user acceptability, with a particular emphasis on perceived
utility and ease of use as crucial factors in the uptake of technology. The

integration of UAT and TAM provided a thorough comprehension of both
123