Page 129 - FULL REPORT 30012024

Figure 4.51 Prediction function codes in flask application

This function precisely converts user inputs into a format that can be

understood by a machine. Subsequently, the encoded variables are
consolidated into a well-organized data frame that conforms to the

specifications of the prediction model. When the function is executed, it
employs a pre-trained machine learning model to calculate the likelihood

of a stroke occurrence. This likelihood is then converted into a percentage

to indicate the user's risk of having a stroke.

Once the ‘predict_stroke_risk’ function is used to calculate the stroke risks,
it takes into account several health and lifestyle parameters including

gender, age, hypertension, heart disease, BMI, smoking status, marital

status, work type, and residence type. The resulting risk score is then
classified as "low," "moderate," or "high." The category is determined by

the numerical value of the stroke risk score. Based on the risk
categorization, the user's BMI is assigned to one of four categories:

"underweight," "normal weight," "overweight," or "obesity." The

categorization is established based on conventional BMI thresholds.

112