The feasibility of bioethanol production using durian seed waste was investigated in this study. The effects of hydrolysis parameters (temperature, time, pressure and solid to water ratio) on the yields of reducing sugars and bioethanol were also examined. Central composite design was used to determine the optimum conditions of both reducing sugars yields obtained from hydrolysis stage and ethanol from reducing sugars fermentation. The optimized values for subcritical water process of durian seeds to produce reducing sugars were achieved at temperature of 139.8°C; solid to water ratio of 1:30; pressure of 30 bar; and reaction time of 3.58 h with 32.37 % yield of reducing sugars. The fermentation of 20 g L -1 reducing sugars for 72 h gave the highest ethanol concentration, i.e., 9.85 g L -1 .

Feature selection in the classification model has a role to choose relevant and interconnected features in the data mining task. in the medical world, feature selection can help the classification model in predicting heart attack. Naive Bayes is one of the most popular classification learning methods that can help to predict patients in helping paramedics to make decisions. The addition of feature selection in the form of backward elimination can increase the accuracy of Naïve Bayes by 89.45% from 84.29% previously. The results of this study indicate the accuracy of a backward selection method in predicting heart attack is quite high in adding accuracy.