1) The document discusses various machine learning and data mining techniques for predicting cardiovascular disease that have been studied in previous literature, including HRFLM, KNN, SVM, ANN, EHR modeling, decision trees, HDFS, and naive Bayes.
2) Accuracy rates from prior studies that used these techniques ranged from 78% to 91.38%. The highest accuracy was obtained using a hybrid of random forest and linear models.
3) Going forward, the authors suggest implementing prediction systems where needed and improving attribute selection to increase accuracy. More research is still needed to better predict early-stage heart disease.