Research and Application of K-means++ Based Clustering Algorithm Combined with C4.5 in the Design of Three-Meal Recipes

With the development of information technology, optimisation algorithms have become increasingly important in the field of diet and nutrition, especially in personalised recipe design. The aim of this study is to apply K-means++ clustering algorithm and C4.5 algorithm to construct a dietary nutrition measurement model for scientific and personalised recipe design to meet individual nutritional needs. In the study, dietary data were preprocessed and discretised using the K-means++ algorithm, followed by the construction of a decision tree model using the C4.5 algorithm, and model validation and optimisation. The optimised C4.5 algorithm improves the accuracy of dietary nutritional classification, especially in under-and over-nutrition, although there is still room for improvement in balanced nutritional classification. The dietary nutrition measurement model combining K-means++ and C4.5 algorithm effectively improves the science and personalisation of recipe design, and provides a reference for healthy diet.