The Causes Analysis of Ischemic Stroke Transformation into Hemorrhagic Stroke using PLS (partial Least Square)-GA and Swarm Algorithm

Ischemic stroke has been known to convert to Hemorrhagic stroke with ease. It is crucial to identify high risk patients with relevant diseases. This study uses the Taiwan National Health Insurance database to collate and analyze the relevant diseases leading to Hemorrhagic from Ischemic stroke. We propose several novel machine learning based algorithms and indexing methods for disease transformation prediction and accuracy enhancements. They are a modified swarm algorithm, partial least square(PLS) algorithm and genetic algorithms. From the petition application files accumulated from 2006 to 2013 within the National Health Insurance Database, 8,483 patients with ischemic stroke is collected. Among them, the 1,145 patients with both ischemic stroke and hemorrhagic stroke were screened according to the ICD-9-CM diagnostic code. The disease history of each patient is vectorized and stacked into a matrix for analysis. The PLS/GA process is then applied on the disease history matrix, trying to filter out the candidate diseases leading to such transformation. A total of 750 diseases were found to be associated with ischemic stroke and hemorrhagic stroke through the PLS/GA process. A modified PSO algorithm is developed to further weight these selected diseases. A quartile rule is then applied to filter these weighted diseases to ten most influential diseases which are dizziness, constipation, chronic renal failure, hypertension, diabetes, hyperlipidemia, anxiety, muscle pain, prostatic hypertrophy, etc. In addition to normally known diseases to such conversion in the literature such as hypertension and diabetes, we also discovered more potential diseases. The initial analysis accuracy of our proposed methods reaches 86% on average, while that of the traditional Neural Network algorithm utilizing the same training data was only 49.5%.