Alternating Projection Algorithm of Linear Regression Method with Convex Constraint in Cerebral Stroke Patients´ Psychological Analysis

Investigations show that the rehabilitation of cerebral stroke patients depends on their psychological condition in particular. Therefore, studying psychological condition of cerebral stroke patients is very important. Psychological characteristics analysis of cerebral stroke patients now becomes the hot domain in biomedical field. In this paper, we propose the key indexes on cerebral stroke patients based on these characteristics, including emotion, depression, and anxiety etc. According to these indexes we build a detailed mental scale. Using the scale we can obtain a norm of cerebral stroke patients. Based on the norm we propose a kind of generalized linear regression model with convex constraint, in which the dependent variable and regression coefficients all are unknown. We give the least square estimation of parameters using the alternating projection algorithm between two convex sets. After getting the solution of the evaluation model we can realize the evaluation cerebral stroke patients. Our work may give a useful tool for evaluation and a precondition for deep research of the cerebral stroke patients.