Sensor Infused Quantum CNN for Diabetes Disease Prediction and Diet Recommendation

Diabetes management requires evidence-based recommendations that enable people to manage their health. A rising diabetes rate can lead to significant health risks and financial hardships. An early diagnosis and efficient treatment are essential to reduce the effects of diabetes. With dietary recommendations and other essential components of diabetes care, complications from diabetes may be reduced, and health can be enhanced. In this paper, a novel Sensor infused QUantum CNN for diabetes Identification and Diet recommendation (SQUID) technique has been proposed, which identifies diabetes using an IoT system and provides diet recommendations for reducing diabetes. The proposed SQUID system collects data from remote patients using IoT sensors and uses the Namib Beetle Optimization (NBO) technique to select the features. The prediction phase uses the Quantum CNN technique for classifying the input into diabetes and non-diabetes. After prediction, the suggestion phase will provide the diet recommendation using the fuzzy rule for the person affected with diabetes through the mobile application. The efficacy of the proposed SQUID framework has been assessed using specific parameters such as Accuracy (AC), Precision (PN), F1 score (F1_S), Recall (RL) and Diagnostic Odds Ratio (DOR). The SQUID framework achieves a higher AC of 98.69%, whereas HCBDA, IWBSOA and e-diagnosis achieve the AC of 92%, 94%, and 96.5%. The proposed SQUID achieved an 89.90% diagnostic odds ratio where the existing HCBDA reached 69.58%, Health-Edge reached 53.86% and IWBSOA obtained a 72.93% diagnostic odds ratio respectively.