Related Experiment Videos
[Experience with the prognosis of strokes]
Insights
This study identifies key risk factors for forecasting ischemic and hemorrhagic strokes using discriminant analysis. The developed prognostic system achieved high accuracy in predicting stroke types.
Area of Science:
- Neurology
- Biostatistics
- Medical Informatics
Abstract:
Results of using discriminant analysis for forecasting ischemic and hemorrhagic strokes are presented. Use was made of clinical data obtained from groups consisting of 400 patients with a history of ischemic strokes, 50 patients with a history of hemorrhagic strokes and 100 healthy individuals. Each observation in these groups was analyzed from 120 parameters and features. As a result of mathematical processing of these data 15 risk factors found to be the most informative for forecasting ischemic strokes, and 16 factors for forecasting hemorrhagic strokes were selected. The reliability of the identification capacity of the prognostic system was evaluated from generalized Machalanobis distance. The correctness of forecasting ischemic strokes was confirmed in 83% and of hemorrhagic strokes in 94.7% of the cases.