A three-step methodology for GI classification, GL estimation of foods by fuzzy c-means classification and fuzzy pattern recognition, and an LP-based diet model for glycaemic control

• Fuzzy c-means classification is used for assigning the foods with measured GI values to GI classes.
• Fuzzy pattern recognition is used for assigning the foods with no measured GI values to GI classes.
• An approach for estimating GL values of foods is proposed.
• Decision rules are developed for menu planning.
• A Linear-Programming-based (LP-based) diet model is proposed for glycaemic control.

In this paper, a three-step methodology is proposed for assigning foods with measured glycaemic index (GI) values to GI classes by using the fuzzy c-means classification technique, assigning foods with no measured GI values to GI classes by using the fuzzy pattern recognition technique, and estimating the glycaemic load (GL) values of foods with no measured GI values. In this methodology, the decision rules for menu planning are also defined, and a Linear Programming-based (LP-based) diet model is developed with the objective function of minimizing the total dietary glycaemic load and the constraints of the daily nutritional requirements. An application based on the real data of GI, GL, and nutritional values of the foods is also provided.