Healthy Diet Food Decision Using Rough-Chi-Squared Goodness

Rough sets approximations have been implemented to handle categorical attributes from various domain problems. However, few studies investigate the combination between dependency degree of rough sets and chi-square test in decision support of the categorical data analysis. In this paper, we are interested to decide the healthy status based on healthy diet foods based on rough sets and chi-square test. To improve the decision accuracy, the data reduction of rough sets is also considered. The gathering data was taken from 20 adult persons using question list form. Results showed that dependency degree and chi-square test after data reduction is better than before. Therefore, the final healthy status (Good) is determined by combination food type, level and dependency degree = {(Food A; High; 50%), (Food; Less; 50%), (Food C; Normal; 44.4%) and (Food F; Less; 27.8%)}. In this case, the accuracy on decision making is really influenced by inconsistent information from the conditional attributes and objects in the data set, thus both can be solved by data reduction strategy. Additionally, the dependency degree is very helpful to support chi-square test for investigating relationship between conditional and decision attributes, especially limited or small data sizes.